模型介绍
IndexTTS 是一个零样本文本转语音（TTS）系统，能够在无需大规模说话人数据的情况下，快速生成高质量、个性化的语音。该模型不仅支持说话人音色的迁移，还提供了灵活的情感与语气控制方式，包括 音频参考、情感向量和文本提示，帮助用户在多样化的应用场景中生成自然且富有表现力的语音。
核心特性：
\uD83C\uDF99 零样本音色克隆：仅需一段短音频即可克隆说话人音色。
\uD83D\uDE03 多维度情感控制：支持通过音频、向量、文本三种方式调节语气。
⚡ 高质量合成：生成语音流畅自然，语气和韵律贴近真实语音。
\uD83D\uDD27 灵活应用：适用于智能客服、游戏配音、虚拟角色、内容创作等场景。
使用方式
模式一：通过音频控制语气
原理：上传一段带有目标情绪的参考音频，模型会模仿其中的语气和韵律。
关键参数：
emo_audio_prompt_url：情感参考音频的地址emo_alpha：情感影响强度（0~1，越大情绪越明显）prompt_audio_url：说话人音色参考音频prompt_text：对应prompt_audio_url的文字内容（用于语义对齐）
适用场景：已有真实录音作为情感样例，想让模型模仿。
模式二：通过文本控制语气
原理：输入一段带有目标情绪的参考文本，模型会根据文字语义推断并模仿对应的语气。
关键参数：
use_emo_text：是否启用情感参考文本（true/false）emo_text：表达目标情绪的参考文本prompt_audio_url：说话人音色参考音频prompt_text：对应prompt_audio_url的文字内容（用于语义对齐）
适用场景：没有情感音频样例，仅通过文字引导模型生成目标情绪语音。
模式三：通过向量控制语气
原理：通过设置情感向量，控制合成语音中不同情绪的占比和强度。
关键参数：
emo_vector：情感强度向量[happy, angry, sad, afraid, disgusted, melancholic, surprised, calm]，数值范围 0~1emo_alpha：情感影响强度（0~1，越大情绪越明显）prompt_audio_url：说话人音色参考音频prompt_text：对应prompt_audio_url的文字内容（用于语义对齐）
适用场景：希望精细调节合成语音的情绪比例，无需真实参考音频。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":238,"name":"语音和音频合成","slug":"text-to-speech","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-to-speech.svg","en_name":"Text To Speech","color":"","description":"","href":"","created_at":1700443954770,"sort_order":0,"categories":[{"id":13,"parent_id":1,"name":"音频","slug":"audio","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/audio.svg","sort_order":13,"en_name":"Audio"},{"id":19,"parent_id":8,"name":"音频","slug":"dataset-audio","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-audio.svg","sort_order":19,"en_name":"Audio"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":23140,"name":"IndexTTS-2","namespace":"hf-models","mirror_namespace":"IndexTeam","mirror_url":"https://huggingface.co/IndexTeam/IndexTTS-2"},"total_calls":8729,"operation_summary":{"service_id":2363,"unit_tag":1261,"min_price":2.0000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":2.0000,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":8},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":444,"account":0,"account_type":0,"type":"text2speech","service_id":2363,"service_type":"serverless","model_id":23140,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"语音生成(音频控制语气)","description":"","path":"v1/audio/speech","protocol":"http","price":2.0000,"average_price":2.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"input","type":"string","control":"text","samples":[{"label":"示例1","value":"我知道自己不是一个人在战斗，有大家的支持和协作，我相信我们一定能一起把事情做好。他的爱像秋天的阳光，看似清冷，却总能在我最需要的时候给予温暖"}],"category":"main","required":true,"description":"合成文本定义了模型要生成的对话内容。"},{"in":"body","name":"prompt_audio_url","type":"string","props":{"enableUrl":true},"control":"file","samples":[{"label":"示例1","value":"https://gitee.com/gitee-ai/moark-assets/raw/master/jay_prompt.wav"}],"category":"main","required":true,"description":"参数用于指定一段在线音频作为参考提示，用来控制合成语音的音色或语气。"},{"in":"body","name":"prompt_text","type":"string","control":"text","samples":[{"label":"示例1","value":"对我来讲是一种荣幸，但是也是压力蛮大的。不过我觉得是一种呃很好的一个挑战。"}],"category":"main","required":false,"description":"参数用于提供与 prompt_audio_url 对应的参考文本，以便模型对齐音频和语义。"},{"in":"body","name":"emo_audio_prompt_url","type":"string","props":{"enableUrl":true},"control":"file","samples":[{"label":"示例1","value":"https://gitee.com/gitee-ai/moark-assets/raw/master/index-tts-2/emo_sad.wav"}],"category":"main","required":true,"description":"参数用于指定一段在线情感音频作为参考提示，用来引导合成语音的情绪和语气表现。"},{"in":"body","name":"emo_alpha","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"参数用于控制情感参考音频对合成语音的影响强度，数值越大，生成语音的情绪特征越明显。","default_value":0.5},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1757675193000,"updated_at":1785400225000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[444],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":543,"account":0,"account_type":0,"type":"text2speech","service_id":0,"model_id":23140,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":1261},{"id":445,"account":0,"account_type":0,"type":"text2speech","service_id":2363,"service_type":"serverless","model_id":23140,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"语音生成(文本控制语气)","description":"","path":"v1/audio/speech","protocol":"http","price":2.0000,"average_price":2.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"input","type":"string","control":"text","samples":[{"label":"示例1","value":"我知道自己不是一个人在战斗，有大家的支持和协作，我相信我们一定能一起把事情做好。他的爱像秋天的阳光，看似清冷，却总能在我最需要的时候给予温暖"}],"category":"main","required":true,"description":"合成文本定义了模型要生成的对话内容。"},{"in":"body","name":"prompt_audio_url","type":"string","props":{"enableUrl":true},"control":"file","samples":[{"label":"示例1","value":"https://gitee.com/gitee-ai/moark-assets/raw/master/jay_prompt.wav"}],"category":"main","required":true,"description":"参数用于指定一段在线音频作为参考提示，用来控制合成语音的音色或语气。"},{"in":"body","name":"prompt_text","type":"string","control":"text","samples":[{"label":"示例1","value":"对我来讲是一种荣幸，但是也是压力蛮大的。不过我觉得是一种呃很好的一个挑战。"}],"category":"main","required":false,"description":"参数用于提供与 prompt_audio_url 对应的参考文本，以便模型对齐音频和语义。"},{"in":"body","name":"emo_text","type":"string","control":"text","samples":[{"label":"示例1","value":"你吓死我了！你是鬼吗？"}],"category":"main","required":false,"description":"用于提供与情感参考音频对应的文本，辅助模型对齐语义和情绪。"},{"in":"body","name":"use_emo_text","type":"boolean","control":"boolean","required":false,"description":"参数用于控制是否启用 emo_text，即是否让模型使用参考文本来对齐情感音频的语义。","default_value":true},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1757675241000,"updated_at":1785400322000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[445],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":544,"account":0,"account_type":0,"type":"text2speech","service_id":0,"model_id":23140,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":1261},{"id":446,"account":0,"account_type":0,"type":"text2speech","service_id":2363,"service_type":"serverless","model_id":23140,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"语音生成异步(音频控制语气)","description":"","path":"v1/async/audio/speech","protocol":"http","price":2.0000,"average_price":2.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"inputs","type":"string","control":"text","samples":[{"label":"示例1","value":"我知道自己不是一个人在战斗，有大家的支持和协作，我相信我们一定能一起把事情做好。他的爱像秋天的阳光，看似清冷，却总能在我最需要的时候给予温暖"}],"category":"main","required":true,"description":"合成文本定义了模型要生成的对话内容。"},{"in":"body","name":"prompt_audio_url","type":"string","props":{"enableUrl":true},"control":"file","samples":[{"label":"示例1","value":"https://gitee.com/gitee-ai/moark-assets/raw/master/jay_prompt.wav"}],"category":"main","required":true,"description":"参数用于指定一段在线音频作为参考提示，用来控制合成语音的音色或语气。"},{"in":"body","name":"prompt_text","type":"string","control":"text","samples":[{"label":"示例1","value":"对我来讲是一种荣幸，但是也是压力蛮大的。不过我觉得是一种呃很好的一个挑战。"}],"category":"main","required":false,"description":"参数用于提供与 prompt_audio_url 对应的参考文本，以便模型对齐音频和语义。"},{"in":"body","name":"emo_audio_prompt_url","type":"string","props":{"enableUrl":true},"control":"file","samples":[{"label":"示例1","value":"https://gitee.com/gitee-ai/moark-assets/raw/master/index-tts-2/emo_sad.wav"}],"category":"main","required":true,"description":"参数用于指定一段在线情感音频作为参考提示，用来引导合成语音的情绪和语气表现。"},{"in":"body","name":"emo_alpha","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"参数用于控制情感参考音频对合成语音的影响强度，数值越大，生成语音的情绪特征越明显。","default_value":0.5},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1757675292000,"updated_at":1785400397000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[446],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":545,"account":0,"account_type":0,"type":"text2speech","service_id":0,"model_id":23140,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":1261},{"id":447,"account":0,"account_type":0,"type":"text2speech","service_id":2363,"service_type":"serverless","model_id":23140,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"语音生成异步(文本控制语气)","description":"","path":"v1/async/audio/speech","protocol":"http","price":2.0000,"average_price":2.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"inputs","type":"string","control":"text","samples":[{"label":"示例1","value":"我知道自己不是一个人在战斗，有大家的支持和协作，我相信我们一定能一起把事情做好。他的爱像秋天的阳光，看似清冷，却总能在我最需要的时候给予温暖"}],"category":"main","required":true,"description":"合成文本定义了模型要生成的对话内容。"},{"in":"body","name":"prompt_audio_url","type":"string","props":{"enableUrl":true},"control":"file","samples":[{"label":"示例1","value":"https://gitee.com/gitee-ai/moark-assets/raw/master/jay_prompt.wav"}],"category":"main","required":true,"description":"参数用于指定一段在线音频作为参考提示，用来控制合成语音的音色或语气。"},{"in":"body","name":"prompt_text","type":"string","control":"text","samples":[{"label":"示例1","value":"对我来讲是一种荣幸，但是也是压力蛮大的。不过我觉得是一种呃很好的一个挑战。"}],"category":"main","required":false,"description":"参数用于提供与 prompt_audio_url 对应的参考文本，以便模型对齐音频和语义。"},{"in":"body","name":"emo_text","type":"string","control":"text","samples":[{"label":"示例1","value":"你吓死我了！你是鬼吗？"}],"category":"main","required":false,"description":"用于提供与情感参考音频对应的文本，辅助模型对齐语义和情绪。"},{"in":"body","name":"use_emo_text","type":"boolean","control":"boolean","required":false,"description":"参数用于控制是否启用 emo_text，即是否让模型使用参考文本来对齐情感音频的语义。","default_value":true},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1757675336000,"updated_at":1785400415000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[447],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":546,"account":0,"account_type":0,"type":"text2speech","service_id":0,"model_id":23140,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":1261},{"id":543,"account":0,"account_type":0,"type":"text2speech","service_id":2363,"service_type":"serverless","model_id":23140,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"语音生成(音频控制语气)","description":"","path":"v1/audio/speech","protocol":"http","price":2.0000,"average_price":2.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"input","type":"string","control":"text","samples":[{"label":"示例1","value":"我知道自己不是一个人在战斗，有大家的支持和协作，我相信我们一定能一起把事情做好。他的爱像秋天的阳光，看似清冷，却总能在我最需要的时候给予温暖"}],"category":"main","required":true,"description":"合成文本定义了模型要生成的对话内容。"},{"in":"body","name":"prompt_audio_url","type":"string","props":{"enableUrl":true},"control":"file","samples":[{"label":"示例1","value":"https://gitee.com/gitee-ai/moark-assets/raw/master/jay_prompt.wav"}],"category":"main","required":true,"description":"参数用于指定一段在线音频作为参考提示，用来控制合成语音的音色或语气。"},{"in":"body","name":"prompt_text","type":"string","control":"text","samples":[{"label":"示例1","value":"对我来讲是一种荣幸，但是也是压力蛮大的。不过我觉得是一种呃很好的一个挑战。"}],"category":"main","required":false,"description":"参数用于提供与 prompt_audio_url 对应的参考文本，以便模型对齐音频和语义。"},{"in":"body","name":"emo_audio_prompt_url","type":"string","props":{"enableUrl":true},"control":"file","samples":[{"label":"示例1","value":"https://gitee.com/gitee-ai/moark-assets/raw/master/index-tts-2/emo_sad.wav"}],"category":"main","required":true,"description":"参数用于指定一段在线情感音频作为参考提示，用来引导合成语音的情绪和语气表现。"},{"in":"body","name":"emo_alpha","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"参数用于控制情感参考音频对合成语音的影响强度，数值越大，生成语音的情绪特征越明显。","default_value":0.5},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1765437402000,"updated_at":1785410657000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[543],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":444,"account":0,"account_type":0,"type":"text2speech","service_id":0,"model_id":23140,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":1261},{"id":544,"account":0,"account_type":0,"type":"text2speech","service_id":2363,"service_type":"serverless","model_id":23140,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"语音生成(文本控制语气)","description":"","path":"v1/audio/speech","protocol":"http","price":2.0000,"average_price":2.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"input","type":"string","control":"text","samples":[{"label":"示例1","value":"我知道自己不是一个人在战斗，有大家的支持和协作，我相信我们一定能一起把事情做好。他的爱像秋天的阳光，看似清冷，却总能在我最需要的时候给予温暖"}],"category":"main","required":true,"description":"合成文本定义了模型要生成的对话内容。"},{"in":"body","name":"prompt_audio_url","type":"string","props":{"enableUrl":true},"control":"file","samples":[{"label":"示例1","value":"https://gitee.com/gitee-ai/moark-assets/raw/master/jay_prompt.wav"}],"category":"main","required":true,"description":"参数用于指定一段在线音频作为参考提示，用来控制合成语音的音色或语气。"},{"in":"body","name":"prompt_text","type":"string","control":"text","samples":[{"label":"示例1","value":"对我来讲是一种荣幸，但是也是压力蛮大的。不过我觉得是一种呃很好的一个挑战。"}],"category":"main","required":false,"description":"参数用于提供与 prompt_audio_url 对应的参考文本，以便模型对齐音频和语义。"},{"in":"body","name":"emo_text","type":"string","control":"text","samples":[{"label":"示例1","value":"你吓死我了！你是鬼吗？"}],"category":"main","required":false,"description":"用于提供与情感参考音频对应的文本，辅助模型对齐语义和情绪。"},{"in":"body","name":"use_emo_text","type":"boolean","control":"boolean","required":false,"description":"参数用于控制是否启用 emo_text，即是否让模型使用参考文本来对齐情感音频的语义。","default_value":true},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1765437507000,"updated_at":1785410668000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[544],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":445,"account":0,"account_type":0,"type":"text2speech","service_id":0,"model_id":23140,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":1261},{"id":545,"account":0,"account_type":0,"type":"text2speech","service_id":2363,"service_type":"serverless","model_id":23140,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"语音生成异步(音频控制语气)","description":"","path":"v1/async/audio/speech","protocol":"http","price":2.0000,"average_price":2.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"inputs","type":"string","control":"text","samples":[{"label":"示例1","value":"我知道自己不是一个人在战斗，有大家的支持和协作，我相信我们一定能一起把事情做好。他的爱像秋天的阳光，看似清冷，却总能在我最需要的时候给予温暖"}],"category":"main","required":true,"description":"合成文本定义了模型要生成的对话内容。"},{"in":"body","name":"prompt_audio_url","type":"string","props":{"enableUrl":true},"control":"file","samples":[{"label":"示例1","value":"https://gitee.com/gitee-ai/moark-assets/raw/master/jay_prompt.wav"}],"category":"main","required":true,"description":"参数用于指定一段在线音频作为参考提示，用来控制合成语音的音色或语气。"},{"in":"body","name":"prompt_text","type":"string","control":"text","samples":[{"label":"示例1","value":"对我来讲是一种荣幸，但是也是压力蛮大的。不过我觉得是一种呃很好的一个挑战。"}],"category":"main","required":false,"description":"参数用于提供与 prompt_audio_url 对应的参考文本，以便模型对齐音频和语义。"},{"in":"body","name":"emo_audio_prompt_url","type":"string","props":{"enableUrl":true},"control":"file","samples":[{"label":"示例1","value":"https://gitee.com/gitee-ai/moark-assets/raw/master/index-tts-2/emo_sad.wav"}],"category":"main","required":true,"description":"参数用于指定一段在线情感音频作为参考提示，用来引导合成语音的情绪和语气表现。"},{"in":"body","name":"emo_alpha","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"参数用于控制情感参考音频对合成语音的影响强度，数值越大，生成语音的情绪特征越明显。","default_value":0.5},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1765437541000,"updated_at":1785410679000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[545],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":446,"account":0,"account_type":0,"type":"text2speech","service_id":0,"model_id":23140,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":1261},{"id":546,"account":0,"account_type":0,"type":"text2speech","service_id":2363,"service_type":"serverless","model_id":23140,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"语音生成异步(文本控制语气)","description":"","path":"v1/async/audio/speech","protocol":"http","price":2.0000,"average_price":2.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"inputs","type":"string","control":"text","samples":[{"label":"示例1","value":"我知道自己不是一个人在战斗，有大家的支持和协作，我相信我们一定能一起把事情做好。他的爱像秋天的阳光，看似清冷，却总能在我最需要的时候给予温暖"}],"category":"main","required":true,"description":"合成文本定义了模型要生成的对话内容。"},{"in":"body","name":"prompt_audio_url","type":"string","props":{"enableUrl":true},"control":"file","samples":[{"label":"示例1","value":"https://gitee.com/gitee-ai/moark-assets/raw/master/jay_prompt.wav"}],"category":"main","required":true,"description":"参数用于指定一段在线音频作为参考提示，用来控制合成语音的音色或语气。"},{"in":"body","name":"prompt_text","type":"string","control":"text","samples":[{"label":"示例1","value":"对我来讲是一种荣幸，但是也是压力蛮大的。不过我觉得是一种呃很好的一个挑战。"}],"category":"main","required":false,"description":"参数用于提供与 prompt_audio_url 对应的参考文本，以便模型对齐音频和语义。"},{"in":"body","name":"emo_text","type":"string","control":"text","samples":[{"label":"示例1","value":"你吓死我了！你是鬼吗？"}],"category":"main","required":false,"description":"用于提供与情感参考音频对应的文本，辅助模型对齐语义和情绪。"},{"in":"body","name":"use_emo_text","type":"boolean","control":"boolean","required":false,"description":"参数用于控制是否启用 emo_text，即是否让模型使用参考文本来对齐情感音频的语义。","default_value":true},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1765437573000,"updated_at":1785410689000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[546],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":447,"account":0,"account_type":0,"type":"text2speech","service_id":0,"model_id":23140,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":1261}]},{"service":{"id":2368,"ident":"Qwen3-Next-80B-A3B-Instruct","type":"serverless","vendor":"","region":"","compute":"","model":23151,"name":"Qwen3-Next-80B-A3B-Instruct","remark":"Qwen3-Next-80B-A3B-Instruct 是 Qwen3-Next 系列的首发大型语言模型，总参数约 80B（激活约 3B），非嵌入参数约 79B 亿采用混合注意力（Gated DeltaNet + Gated Attention）与高稀疏度 Mixture-of-Experts（MoE）架构，面向指令遵循与通用对话场景，兼顾参数效率与推理吞吐。","price":0.0000,"props":{"name":"Qwen3-Next-80B-A3B-Instruct","service_type":"serverless"},"inventories":0,"created_at":1757734770000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3-Next-80B-A3B-Instruct 是 Qwen3-Next 系列的首发大型语言模型，总参数约 80B（激活约 3B），非嵌入参数约 79B 亿采用混合注意力（Gated DeltaNet + Gated Attention）与高稀疏度 Mixture-of-Experts（MoE）架构，面向指令遵循与通用对话场景，兼顾参数效率与推理吞吐。相比同代模型，其在知识、推理、编码、多语言与代理等多项基准上表现强劲。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":52,"name":"Apache license 2.0","slug":"apache-2.0","icon":"","en_name":"","color":"","description":"","href":"","created_at":1700443952113,"sort_order":0,"categories":[{"id":4,"parent_id":0,"name":"许可证","slug":"license","created_at":1700107824458,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/license.svg","sort_order":4,"en_name":""}]},{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":10,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1150,"name":"32K","slug":"context_size_32k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-f8crosrdk-id-y4z8cy4z9-context_size.png","en_name":"32K","color":"blue","description":"上下文大小","href":"","created_at":1734693847615,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1219,"name":"128K","slug":"context_size_128k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-zd1snxypr-context_size.png","en_name":"128K","color":"blue","description":"上下文大小","href":"","created_at":1747894585219,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":23151,"name":"Qwen3-Next-80B-A3B-Instruct","namespace":"hf-models","mirror_namespace":"Qwen","mirror_url":"https://huggingface.co/Qwen/Qwen3-Next-80B-A3B-Instruct"},"total_calls":8278,"operation_summary":{"service_id":2368,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":1.00000000,"min_output_million_tokens_price":4.00000000,"max_price":0.0200,"max_input_million_tokens_price":1.00000000,"max_output_million_tokens_price":4.00000000,"free_operation_count":1,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":659,"account":0,"account_type":0,"type":"text2text","service_id":2368,"service_type":"serverless","model_id":23151,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0200,"average_price":0.0200,"input_million_tokens_price":1.00000000,"output_million_tokens_price":4.00000000,"input_million_tokens_price_average":1.00000000,"output_million_tokens_price_average":4.00000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1774000448000,"updated_at":1775023402000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[659],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":710,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":23151,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":710,"account":0,"account_type":0,"type":"text2text","service_id":2368,"service_type":"serverless","model_id":23151,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":1.00000000,"output_million_tokens_price":4.00000000,"input_million_tokens_price_average":1.00000000,"output_million_tokens_price_average":4.00000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1776911278000,"updated_at":1786965040000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[710],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":659,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":23151,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":2436,"ident":"MinerU2.5","type":"serverless","vendor":"","region":"","compute":"","model":23170,"name":"MinerU 2.5","remark":"MinerU 2.5-2509-1.2B 是一款约 11 亿参数的开源视觉语言模型，采用 AGPL-3.0 许可，专注于 OCR 与复杂文档解析，能够高效提取图像中的文字和结构化信息，适用于发票、合同等多样化文档的自动化处理。","price":0.0000,"props":{"name":"MinerU2.5-2509-1.2B","service_type":"serverless"},"inventories":0,"created_at":1758695057000,"updated_at":1759142056000,"sort_order":0,"status":1,"recommend":true,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-btfubx0rb-avatar-mineru.png","description_html":"简介
名称：MinerU2.5-2509-1.2B
发布者：OpenDataLab
参数规模：约 1.16B（大约 11.6 亿参数）
License（许可）：AGPL-3.0
功能与定位
MinerU2.5 是一个视觉-语言（Vision-Language）模型，专门用于 OCR（光学字符识别）和文档解析任务，尤其是复杂多样的现实世界文档。
它的设计目标包括：
从图像中识别文字（OCR）
将文档中的结构或内容块（text blocks / layout blocks）提取出来进行解析，比如段落、表格、标题、正文等
能够处理多样化格式与布局，不仅限于简单排版的文档，也包括带有复杂版式或者视觉噪声的真实场景文档
\uD83E\uDDE0 Qwen3-VL-30B-A3B-Instruct 模型介绍
Qwen3-VL-30B-A3B-Instruct 是 Qwen 团队推出的 旗舰级多模态指令大模型（Vision–Language Large Model），
具备卓越的 视觉理解、跨模态推理与时序感知能力。
相比 4B 与 8B 版本，30B-A3B-Instruct 拥有更高的参数规模与更深层次的特征建模能力，
可对图像、视频及文本进行统一理解与自然语言生成。
该模型特别强化了 视觉语义融合 与 时间维度建模，能实现从静态图像到动态视频的复杂内容理解，
并在多轮指令对话、逻辑推理及多场景问答中表现出色，是 Qwen3-VL 系列的旗舰代表。
\uD83E\uDDE9 模型核心特点
跨模态理解与推理
同时理解图像、文字乃至视频帧序列，具备强大的视觉语义整合能力。
可在复杂多模态输入中完成逻辑分析与深层推理，如动作推断、事件因果关系判断等。
视觉识别与时序感知能力强化
采用 A3B 混合架构（Aligned–Adaptive–Auto-Balance），实现视觉–语言特征动态均衡。
支持时间维度建模（Temporal Alignment），能理解视频画面间的事件演化与时序变化。
强大的OCR与视觉识别性能
支持识别多语言文本（含中英日韩等 30+ 语言），在模糊、倾斜、低光环境下仍保持高精度。
能识别多种视觉元素：人物、商品、地标、图表、动植物等。
增强的语言生成与指令跟随能力
“Instruct” 版本经过强化训练，响应精准，能针对复杂任务生成结构化、逻辑化输出。
适合图文结合类智能问答、报告生成、教学讲解等需要语义连贯的应用。
\uD83D\uDE80 典型使用场景
场景类型 | 说明 | 示例 |
\uD83C\uDFAC 视频理解与时序推理 | 理解视频片段中的事件顺序、行为逻辑和情绪变化 | “总结这段视频的主要内容及时间顺序。”、“分析人物的行为变化。” |
\uD83D\uDDBC 复杂图像问答与解释 | 从图像中提取深层语义信息并进行逻辑回答 | “图片中两个人的关系是什么？”、“根据场景判断这可能是哪种活动？” |
\uD83D\uDCCA 多模态分析与报告生成 | 结合图像/文字生成摘要、结论或结构化报告 | “为这张销售趋势图生成一段分析说明。”、“生成一份图文组合报告。” |
\uD83E\uDDFE OCR 信息提取与理解 | 自动读取图片中的文字、表格与文档信息 | “提取并解释截图中的表格数据。”、“识别发票图片内容。” |
\uD83C\uDF93 教育与科研辅助 | 讲解实验过程、分析图表、解释视频教学内容 | “解释实验录像中的关键步骤。”、“根据图表说明温度变化规律。” |
\uD83E\uDD16 智能视觉助手与多模态交互 | 支持图像、视频与文本混合输入的自然对话交互 | “根据上传视频生成教学脚本。”、“结合图片与说明判断系统错误原因。” |
Qwen3Guard是基于Qwen3模型构建的全新安全审查模型系列，旨在为全球用户提供实时、高效的人工智能安全保障。
核心技术：
卓越的语言支持：Qwen3Guard能够支持多达119种语言和方言，满足全球化应用场景的需求，确保不同文化和地区的用户都能获得同样水平的安全保护。
灵活的模型尺寸选择：为了适应不同的部署环境和性能要求，该系列提供了三种不同规模的模型：分别是Qwen3Guard-Gen-0.6B、Qwen3Guard-Gen-4B和Qwen3Guard-Gen-8B，方便开发者根据具体需求进行选择。
强大的上下文安全分析：Qwen3Guard-Gen模型则专注于稳健的全上下文安全分析，尤其适合作为强化学习（RL）中的奖励模型，从更深层次提升AI模型的安全性。
精细化的三级风险分类：模型采用三级风险分类系统，将内容明确划分为“安全”、“争议性”和“不安全”，为内容审查提供了清晰、直观的判断依据。
行业领先的性能表现：在包括英语和中文在内的多项安全基准测试中，Qwen3Guard均展现了当前最顶尖（SOTA）的性能，证明了其在识别和管理不安全内容方面的卓越能力。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":52,"name":"Apache license 2.0","slug":"apache-2.0","icon":"","en_name":"","color":"","description":"","href":"","created_at":1700443952113,"sort_order":0,"categories":[{"id":4,"parent_id":0,"name":"许可证","slug":"license","created_at":1700107824458,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/license.svg","sort_order":4,"en_name":""}]},{"id":1150,"name":"32K","slug":"context_size_32k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-f8crosrdk-id-y4z8cy4z9-context_size.png","en_name":"32K","color":"blue","description":"上下文大小","href":"","created_at":1734693847615,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1176,"name":"风控识别","slug":"risk_control_identification_label","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-risk_control_identification_label.svg","en_name":"Risk Control Identification","color":"","description":"","href":"","created_at":1736506883327,"sort_order":2,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":23187,"name":"Qwen3Guard-Gen-8B","namespace":"hf-models","mirror_namespace":"Qwen","mirror_url":"https://huggingface.co/Qwen/Qwen3Guard-Gen-8B"},"total_calls":66,"operation_summary":{"service_id":2566,"unit_tag":0,"min_price":0.0003,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0003,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":468,"account":0,"account_type":0,"type":"text2text","service_id":2566,"service_type":"serverless","model_id":23187,"vendor":"biren-tech","region":"prod-br-sh-wz-ip","compute":"biren-tech/prod-br-sh-wz-ip/serverless_api/gpu/large/contest","name":"文本风控识别","description":"","path":"v1/chat/completions","protocol":"http","price":0.0003,"average_price":0.0003,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","billing_policy":{},"sort_order":0,"params":[],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1760601415000,"updated_at":1788406450000,"vendor_info":{"id":0,"name":"biren-tech","label":"壁仞科技","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-5rdewrmcj-cover-biren.png"},"region_info":{"name":"prod-br-sh-wz-ip","label":"壁仞上海和乌镇","props":{}},"compute_info":{"id":"biren-tech/prod-br-sh-wz-ip/serverless_api/gpu/large/contest","instance_size":"large","instance_type":"contest","price_per_hour":3.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"壁砺系列","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-br-sh-wz-ip","vendor_name":"biren-tech","type":"SERVERLESS_API","storage":0},"ids":[468],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":2567,"ident":"Qwen3Guard-Gen-4B","type":"serverless","vendor":"","region":"","compute":"","model":23188,"name":"Qwen3Guard-Gen-4B","remark":"Qwen3Guard是阿里巴巴Qwen团队于2025年9月24日发布的人工智能安全审查模型系列。该模型基于Qwen3模型构建，主要用于提供实时高效的人工智能安全保障。","price":0.0000,"props":{"name":"Qwen3Guard-Gen-4B","service_type":"serverless"},"inventories":0,"created_at":1760601581000,"updated_at":1760603605000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3Guard是基于Qwen3模型构建的全新安全审查模型系列，旨在为全球用户提供实时、高效的人工智能安全保障。
核心技术：
卓越的语言支持：Qwen3Guard能够支持多达119种语言和方言，满足全球化应用场景的需求，确保不同文化和地区的用户都能获得同样水平的安全保护。
灵活的模型尺寸选择：为了适应不同的部署环境和性能要求，该系列提供了三种不同规模的模型：分别是Qwen3Guard-Gen-0.6B、Qwen3Guard-Gen-4B和Qwen3Guard-Gen-8B，方便开发者根据具体需求进行选择。
强大的上下文安全分析：Qwen3Guard-Gen模型则专注于稳健的全上下文安全分析，尤其适合作为强化学习（RL）中的奖励模型，从更深层次提升AI模型的安全性。
精细化的三级风险分类：模型采用三级风险分类系统，将内容明确划分为“安全”、“争议性”和“不安全”，为内容审查提供了清晰、直观的判断依据。
行业领先的性能表现：在包括英语和中文在内的多项安全基准测试中，Qwen3Guard均展现了当前最顶尖（SOTA）的性能，证明了其在识别和管理不安全内容方面的卓越能力。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":52,"name":"Apache license 2.0","slug":"apache-2.0","icon":"","en_name":"","color":"","description":"","href":"","created_at":1700443952113,"sort_order":0,"categories":[{"id":4,"parent_id":0,"name":"许可证","slug":"license","created_at":1700107824458,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/license.svg","sort_order":4,"en_name":""}]},{"id":1150,"name":"32K","slug":"context_size_32k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-f8crosrdk-id-y4z8cy4z9-context_size.png","en_name":"32K","color":"blue","description":"上下文大小","href":"","created_at":1734693847615,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1176,"name":"风控识别","slug":"risk_control_identification_label","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-risk_control_identification_label.svg","en_name":"Risk Control Identification","color":"","description":"","href":"","created_at":1736506883327,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":23188,"name":"Qwen3Guard-Gen-4B","namespace":"hf-models","mirror_namespace":"Qwen","mirror_url":"https://huggingface.co/Qwen/Qwen3Guard-Gen-4B"},"total_calls":49,"operation_summary":{"service_id":2567,"unit_tag":0,"min_price":0.0002,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0002,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":469,"account":0,"account_type":0,"type":"text2text","service_id":2567,"service_type":"serverless","model_id":23188,"vendor":"biren-tech","region":"prod-br-sh-wz-ip","compute":"biren-tech/prod-br-sh-wz-ip/serverless_api/gpu/large/contest","name":"文本风控识别","description":"","path":"v1/chat/completions","protocol":"http","price":0.0002,"average_price":0.0002,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","billing_policy":{},"sort_order":0,"params":[],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1760601704000,"updated_at":1788406436000,"vendor_info":{"id":0,"name":"biren-tech","label":"壁仞科技","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-5rdewrmcj-cover-biren.png"},"region_info":{"name":"prod-br-sh-wz-ip","label":"壁仞上海和乌镇","props":{}},"compute_info":{"id":"biren-tech/prod-br-sh-wz-ip/serverless_api/gpu/large/contest","instance_size":"large","instance_type":"contest","price_per_hour":3.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"壁砺系列","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-br-sh-wz-ip","vendor_name":"biren-tech","type":"SERVERLESS_API","storage":0},"ids":[469],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":2568,"ident":"Qwen3Guard-Gen-0.6B","type":"serverless","vendor":"","region":"","compute":"","model":23182,"name":"Qwen3Guard-Gen-0.6B","remark":"Qwen3Guard是阿里巴巴Qwen团队于2025年9月24日发布的人工智能安全审查模型系列。该模型基于Qwen3模型构建，主要用于提供实时高效的人工智能安全保障。","price":0.0000,"props":{"name":"Qwen3Guard-Gen-0.6B","service_type":"serverless"},"inventories":0,"created_at":1760601889000,"updated_at":1760603580000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3Guard是基于Qwen3模型构建的全新安全审查模型系列，旨在为全球用户提供实时、高效的人工智能安全保障。
核心技术：
卓越的语言支持：Qwen3Guard能够支持多达119种语言和方言，满足全球化应用场景的需求，确保不同文化和地区的用户都能获得同样水平的安全保护。
灵活的模型尺寸选择：为了适应不同的部署环境和性能要求，该系列提供了三种不同规模的模型：分别是Qwen3Guard-Gen-0.6B、Qwen3Guard-Gen-4B和Qwen3Guard-Gen-8B，方便开发者根据具体需求进行选择。
强大的上下文安全分析：Qwen3Guard-Gen模型则专注于稳健的全上下文安全分析，尤其适合作为强化学习（RL）中的奖励模型，从更深层次提升AI模型的安全性。
精细化的三级风险分类：模型采用三级风险分类系统，将内容明确划分为“安全”、“争议性”和“不安全”，为内容审查提供了清晰、直观的判断依据。
行业领先的性能表现：在包括英语和中文在内的多项安全基准测试中，Qwen3Guard均展现了当前最顶尖（SOTA）的性能，证明了其在识别和管理不安全内容方面的卓越能力。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":52,"name":"Apache license 2.0","slug":"apache-2.0","icon":"","en_name":"","color":"","description":"","href":"","created_at":1700443952113,"sort_order":0,"categories":[{"id":4,"parent_id":0,"name":"许可证","slug":"license","created_at":1700107824458,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/license.svg","sort_order":4,"en_name":""}]},{"id":1150,"name":"32K","slug":"context_size_32k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-f8crosrdk-id-y4z8cy4z9-context_size.png","en_name":"32K","color":"blue","description":"上下文大小","href":"","created_at":1734693847615,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1176,"name":"风控识别","slug":"risk_control_identification_label","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-risk_control_identification_label.svg","en_name":"Risk Control Identification","color":"","description":"","href":"","created_at":1736506883327,"sort_order":1,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":23182,"name":"Qwen3Guard-Gen-0.6B","namespace":"hf-models","mirror_namespace":"Qwen","mirror_url":"https://huggingface.co/Qwen/Qwen3Guard-Gen-0.6B"},"total_calls":47,"operation_summary":{"service_id":2568,"unit_tag":0,"min_price":0.0001,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0001,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":470,"account":0,"account_type":0,"type":"text2text","service_id":2568,"service_type":"serverless","model_id":23182,"vendor":"biren-tech","region":"prod-br-sh-wz-ip","compute":"biren-tech/prod-br-sh-wz-ip/serverless_api/gpu/large/contest","name":"文本风控识别","description":"","path":"v1/chat/completions","protocol":"http","price":0.0001,"average_price":0.0001,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","billing_policy":{},"sort_order":0,"params":[],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1760601950000,"updated_at":1788406350000,"vendor_info":{"id":0,"name":"biren-tech","label":"壁仞科技","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-5rdewrmcj-cover-biren.png"},"region_info":{"name":"prod-br-sh-wz-ip","label":"壁仞上海和乌镇","props":{}},"compute_info":{"id":"biren-tech/prod-br-sh-wz-ip/serverless_api/gpu/large/contest","instance_size":"large","instance_type":"contest","price_per_hour":3.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"壁砺系列","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-br-sh-wz-ip","vendor_name":"biren-tech","type":"SERVERLESS_API","storage":0},"ids":[470],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":2571,"ident":"AudioFly","type":"serverless","vendor":"","region":"","compute":"","model":23224,"name":"AudioFly","remark":"AudioFly模型是科大讯飞推出的，可基于文本描述输入，生成44.1kHz 采样率的高质量音频，在文本与音效的匹配度上表现优异。AudioFly采用了潜在扩散模型（LDM）架构，经海量多元声音数据训练，支持单事件、多事件场景。","price":0.0000,"props":{"name":"AudioFly","service_type":"serverless"},"inventories":0,"created_at":1760610592000,"updated_at":1760665266000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-q0kpp34ne-avatar-iflytek.png","description_html":"AudioFly模型是科大讯飞推出的，可基于文本描述输入，生成44.1kHz 采样率的高质量音频，在文本与音效的匹配度上表现优异。AudioFly采用了潜在扩散模型（LDM）架构，经海量多元声音数据训练，支持单事件、多事件场景。
在未进行针对性微调的情况下，AudioFly在学术数据集AudioCaps测试集上取得了SOTA指标，证明了该模型良好的推广性和鲁棒性。
![\"\"](\"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-lg56kh1mb-640.webp\" "\\"\\"")
* 以上为在学界大规模音频事件描述数据集AudioCaps上的评测结果
FD：衡量生成音频与真实音频特征分布上的差异
KL：衡量生成结果与真实数据概率分布上的差异
CLAP：计算生成的音频与文本描述的一致性
* 基线模型的指标引用了论文 Stable Audio Open 中的评测分数；AudioFly 评测复用 Stable Audio Open 开源的测试方法保证客观性。
AudioFly的开源能够有效降低音效生成模型的使用门槛，开发者可以将其应用于短视频配音、有声故事生成等领域，释放更多声音创意潜力，创造出更多优质的声音内容。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":52,"name":"Apache license 2.0","slug":"apache-2.0","icon":"","en_name":"","color":"","description":"","href":"","created_at":1700443952113,"sort_order":0,"categories":[{"id":4,"parent_id":0,"name":"许可证","slug":"license","created_at":1700107824458,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/license.svg","sort_order":4,"en_name":""}]},{"id":238,"name":"语音和音频合成","slug":"text-to-speech","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-to-speech.svg","en_name":"Text To Speech","color":"","description":"","href":"","created_at":1700443954770,"sort_order":0,"categories":[{"id":13,"parent_id":1,"name":"音频","slug":"audio","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/audio.svg","sort_order":13,"en_name":"Audio"},{"id":19,"parent_id":8,"name":"音频","slug":"dataset-audio","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-audio.svg","sort_order":19,"en_name":"Audio"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1174,"name":"英文","slug":"en_language_label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735647874656,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":23224,"name":"AudioFly","namespace":"iflytekopensource","title":"AudioFly"},"total_calls":28,"operation_summary":{"service_id":2571,"unit_tag":0,"min_price":0.0200,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0200,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":474,"account":0,"account_type":0,"type":"text2speech","service_id":2571,"service_type":"serverless","model_id":23224,"vendor":"iluvatar","region":"prod-il-sh-if","compute":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","name":"音效生成异步","description":"","path":"v1/async/audio/speech","protocol":"http","price":0.0200,"average_price":0.0200,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"inputs","type":"string","control":"text","samples":[{"label":"猫叫声","value":"A cat meows, and it's drizzling"},{"label":"打雷声","value":"thunder rumbled in the distance"}],"category":"main","required":true,"description":"生成内容"},{"in":"body","name":"num_inference_steps","step":1,"type":"number","control":"slider","maximum":200,"minimum":1,"required":false,"description":"表示模型在生成过程中执行的去噪迭代步数——步数越多，生成结果越精细但速度越慢。","default_value":200},{"in":"body","name":"guidance_scale","step":0.5,"type":"number","control":"slider","maximum":10,"minimum":1,"required":false,"description":"示模型生成内容时对文本提示的依从强度——值越大越贴合提示，越小则更自由多样。","default_value":3.5},{"in":"form","name":"output_format","type":"string","control":"select","options":[{"label":"mp3","value":"mp3"},{"label":"wav","value":"wav"}],"required":false,"description":"输出格式","default_value":""}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1760612063000,"updated_at":1782714142000,"vendor_info":{"id":0,"name":"iluvatar","label":"天数智芯","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-lod1uv2gx-Group 62.svg"},"region_info":{"name":"prod-il-sh-if","label":"上海","props":{"products":["APP","ENDPOINT"]}},"compute_info":{"id":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","instance_size":"2.large","instance_type":"mr-v100","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":2,"cpus_num":12,"memory_gb":32,"architecture":"智铠 100","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-il-sh-if","vendor_name":"iluvatar","type":"SERVERLESS_API","storage":0},"ids":[474],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":2627,"ident":"DeepSeek-OCR","type":"serverless","vendor":"","region":"","compute":"","model":23453,"name":"DeepSeek-OCR","remark":"DeepSeek-OCR 是由 DeepSeek 团队 开发的高精度光学字符识别（OCR, Optical Character Recognition）系统，旨在在复杂场景下实现高效、准确的文字检测与识别。\n它支持 多语言识别、复杂版面分析、表格与公式识别 等多种任务，适用于票据、文档、截图、自然场景图片等多样化输入。","price":0.0000,"props":{"name":"DeepSeek-OCR","service_type":"serverless"},"inventories":0,"created_at":1761128296000,"updated_at":1786809584000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-00z769y98-deepseek.png","description_html":"\uD83E\uDDE0 DeepSeek-OCR 简介
DeepSeek-OCR 是由 DeepSeek 团队 开发的高精度光学字符识别（OCR, Optical Character Recognition）系统，旨在在复杂场景下实现高效、准确的文字检测与识别。
它支持 多语言识别、复杂版面分析、表格与公式识别 等多种任务，适用于票据、文档、截图、自然场景图片等多样化输入。
\uD83D\uDD0D 核心特点
多语言支持：支持中、英、日、韩等多语种文字识别。
高精度识别：通过深度学习视觉语言模型，有效处理噪声、倾斜、模糊等复杂情况。
端到端识别：集成文字检测（Text Detection）与文字识别（Text Recognition）两阶段任务。
轻量高效：支持 GPU 加速与 CPU 推理，可部署在服务器或本地端设备。
结构化输出：支持输出带坐标的文本框信息，便于版面分析或区域提取。
\uD83D\uDE80 应用场景
发票、合同、票据识别
扫描文档数字化
图片翻译与搜索
AI 知识抽取与文本挖掘
智能表单与档案系统
模型简介
SeedVR2-3B 是 ByteDance-Seed 推出的 高性能图像修复与画质增强模型，能够对低质量图片进行智能修复，包括提升清晰度、去除噪点、去压缩伪影、改善模糊、增强纹理细节等。
通过轻量化设计与高效推理，SeedVR2-3B 非常适合接入各类图像处理业务，支持对普通拍摄图、用户上传图、历史照片等多类型图像进行质量提升。
开发者通过调用你们的 API，只需提供原始图像，模型即可自动输出更清晰、更真实的修复图，无需额外复杂配置。
核心能力
✔ 清晰度增强（Enhancement）
提升图像的整体锐度，让人物或物体边缘更清晰、细节更突出。
✔ 图像降噪（Denoise）
有效去除噪点和颗粒感，对低光画像、手机夜拍、老旧照片都有明显改善。
✔ 去压缩伪影（Artifacts Removal）
消除 JPEG 压缩导致的马赛克、方块噪声和涂抹痕迹，使图像更自然。
✔ 模糊修复（Deblur）
改善轻度对焦不准、抖动带来的模糊，让图像更加清晰。
✔ 纹理细节增强（Texture Enhancement）
恢复或增强物体的表面纹理，让画面层次更丰富。
使用场景
使用场景 | 说明 |
UGC / 内容平台图片增强 | 自动提升用户上传图片的清晰度，减少噪点与压缩痕迹，提升社区内容质量。 |
电商商品图优化 | 增强商品展示图的细节和质感，使产品展示更专业，有助于提高转化率。 |
老照片 / 扫描件修复 | 对历史照片、扫描件提升清晰度、去噪、修复模糊，恢复更多细节。 |
人像清晰化处理 | 优化头像、人脸区域的细节表现，使人像更清晰自然。 |
智能相册与相片管理 | 自动清洗图库中的低质量图片，统一画质，提升相册产品使用体验。 |
在线修图工具 / 编辑器 | 为编辑器提供一键画质提升功能，改善拍摄不佳的图像。 |
HunyuanVideo-1.5 是腾讯开源的轻量级视频生成模型，模型在时序一致性、细节表现、运动自然度以及提示词理解方面具有强大的能力，适合多种视频生成与创意场景。
⚙ 模型能力（Capabilities）
1. 文本生成视频（Text-to-Video，T2V）
根据自然语言描述生成高质量短视频
支持人物、场景、物体、动作等多类型生成
具备较好的提示词理解能力，可控制风格、镜头、构图等细节
输出视频稳定、连贯，帧间抖动低
2. 强时序一致性（Temporal Consistency）
视频人物、物体、光影随时间变化自然稳定
大幅减少抖动、面部变化、物体消失等常见问题
3. 高质量生成能力
生成分辨率高、细节丰富
支持多种风格：写实、二次元、电影风、时尚风、艺术风
能表达较复杂的动态动作（如旋转、跳跃、跑动等）
\uD83D\uDCCC 应用场景（Use Cases）
以下以表格形式呈现，方便直接放入你的 API 文档。
应用场景 | 说明 |
内容创作 / 自媒体视频生成 | 根据文本自动生成短视频素材，帮助创作者快速制作内容。 |
角色动画 & 虚拟人生成 | 生成人物动态、表情变化、走秀等动作场景，用于虚拟主播、角色演示。 |
广告与营销素材制作 | 快速生成宣传视频、产品动态展示、创意场景动画。 |
游戏与影视概念设计 | 通过文本或图像生成预览镜头，用于分镜构思和场景预演。 |
教育与科普视频生成 | 用文本描述构造教学场景、流程演示或抽象概念动态可视化。 |
AIGC 创意实验 | 用于生成实验性艺术效果、变换风格、奇幻场景。 |
文心新一代模型文心5.0是原生全模态大模型，采用原生的全模态统一建模技术，将文本、图像、音频、视频联合建模，具备综合的全模态能力。文心5.0基础能力全面升级，基准测试集表现出色，多模态理解、指令遵循、创意写作、事实性、智能体规划与工具应用等表现尤其出色。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1219,"name":"128K","slug":"context_size_128k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-zd1snxypr-context_size.png","en_name":"128K","color":"blue","description":"上下文大小","href":"","created_at":1747894585219,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1220,"name":"64K","slug":"max_tokens_64k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-xroamij61-context_size (1).png","en_name":"64K","color":"blue","description":"上下文长度","href":"","created_at":1748699035394,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":23983,"name":"ERNIE-4.5-0.3B-Paddle","namespace":"hf-models","mirror_namespace":"baidu","mirror_url":"https://huggingface.co/baidu/ERNIE-4.5-0.3B-Paddle"},"total_calls":34,"operation_summary":{"service_id":3881,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":4.00000000,"min_output_million_tokens_price":16.00000000,"max_price":0.0000,"max_input_million_tokens_price":4.00000000,"max_output_million_tokens_price":16.00000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":516,"account":0,"account_type":0,"type":"text2text","service_id":3881,"service_type":"serverless","model_id":23983,"vendor":"baidu-cloud","region":"prod-bce-default-ip","compute":"baidu-cloud/prod-bce-default-ip/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":4.00000000,"output_million_tokens_price":16.00000000,"input_million_tokens_price_average":4.00000000,"output_million_tokens_price_average":16.00000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":512},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":1},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":50,"minimum":0,"required":false,"description":"取值范围为 [0,50]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":1},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":0},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"文件类型：.jpeg, .jpg, .png, .webp，最大体积：5MB"},{"in":"body","name":"messages","type":"string","control":"string","samples":[{"label":"宇宙飞船","value":[{"role":"user","content":[{"text":"这张图的内容是什么？","type":"text"},{"type":"image_url","image_url":{"url":"https://ai-assets.gitee.com/landing-page/example1.png"}}]}]},{"label":"中文诗","value":[{"role":"user","content":[{"text":"图中文字是什么？","type":"text"},{"type":"image_url","image_url":{"url":"https://ai-assets.gitee.com/landing-page/example2.png"}}]}]},{"label":"英文","value":[{"role":"user","content":[{"text":"图片中内容是什么，介绍它的出处，并翻译成中文","type":"text"},{"type":"image_url","image_url":{"url":"https://ai-assets.gitee.com/landing-page/example3.png"}}]}]}],"category":"main","required":true},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1764578786000,"updated_at":1786965473000,"vendor_info":{"id":0,"name":"baidu-cloud","label":"百度智能云","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-w36p9kcyq-3412a51e.png"},"region_info":{"name":"prod-bce-default-ip","label":"百度智能云默认","props":{}},"compute_info":{"id":"baidu-cloud/prod-bce-default-ip/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"prod-bce-default-ip","vendor_name":"baidu-cloud","type":"SERVERLESS_API","storage":0},"ids":[516],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":3973,"ident":"z-image-turbo","type":"serverless","vendor":"","region":"","compute":"","model":23857,"name":"z-image-turbo","remark":"Z-Image-Turbo 是 Z-Image 系列的一员 — 一个高效、轻量级的图像生成基础模型（text-to-image），经过优化与蒸馏 (distillation) 后的 “Turbo” 版本。它旨在以较低的资源成本 (consumer-grade 单卡 < 16GB VRAM) 提供高质量、实用性强的图像生成能力。","price":0.0000,"props":{"name":"Z-Image-Turbo","service_type":"serverless"},"inventories":0,"created_at":1764674347000,"updated_at":1764754807000,"sort_order":0,"status":1,"recommend":true,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"\uD83D\uDD25 Z-Image-Turbo 简介
Z-Image-Turbo 是 Z-Image 系列的一员 — 一个高效、轻量级的图像生成基础模型（text-to-image），经过优化与蒸馏 (distillation) 后的 “Turbo” 版本。它旨在以较低的资源成本 (consumer-grade 单卡 < 16GB VRAM) 提供高质量、实用性强的图像生成能力。
⚙ 基本信息
参数量：约 6 亿参数 (6B)
采样步数 (Inference Steps)：8 步 (8 NFEs) —— 即 “Turbo 模式”
硬件要求：支持在 16GB 显存或更低显卡上运行，适合大多数消费级 GPU
许可协议：开源 (Apache-2.0)，适合商业及个人用途
✅ 核心能力与优势
• 快速 & 高效
相较于传统需要几十步 (50–100 步) 的扩散模型，Z-Image-Turbo 仅需 8 步即可生成高质量图像 → 极大降低延迟与资源消耗。
对于商业级 GPU, 推理延迟可达到亚秒级；对消费级显卡也能保持流畅生成体验。
• 高画质 & 写实能力
能够生成具有“照片级”真实感 (photorealistic) 的场景、人物、产品等图像 — 光影、材质、构图、细节表现出色。
对构图、纹理、环境光等均有良好表现，适用于从艺术创作到商业图像生成的多种需求。
• 双语 & 文本渲染能力
支持中英文提示词 (prompt) —— 无论是生成场景、海报、包含文字的画面 (如广告、UI mockup、包装) 都能准确渲染对应语言文本。
对中文 prompt 的兼容性优秀，适合中国及全球双语环境。
• 资源门槛低 & 可广泛部署
模型轻量 (6B) + 低显存要求，使其可部署在消费级 GPU、本地机器、轻量服务器或云端。
对于个人创作者、小团队、初创公司来说，是“高性能 + 低成本”的理想选型。
• 开源 & 社区友好
模型权重和代码公开 (Apache-2.0)，可以自由使用、部署、商用。
支持多种推理框架、兼容常见生成 / 编辑 / UI 流程，便于集成与二次开发。
DeepSeek-V3.2 是一个在高计算效率与卓越推理和代理性能之间取得平衡的模型。
DeepSeek-V3.2 相比之前版本，对聊天模板进行了重大更新。主要变化包括工具调用格式的修订以及引入“工具思维”功能。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":36,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1219,"name":"128K","slug":"context_size_128k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-zd1snxypr-context_size.png","en_name":"128K","color":"blue","description":"上下文大小","href":"","created_at":1747894585219,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1273,"name":"8K","slug":"context_size_output_8k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-xgg1k2as1-arrow-down.png","en_name":"8K","color":"blue","description":"最大生成长度","href":"","created_at":1770263067993,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":24012,"name":"DeepSeek-V3.2","namespace":"hf-models","mirror_namespace":"deepseek-ai","mirror_url":"https://huggingface.co/deepseek-ai/DeepSeek-V3.2"},"total_calls":9515,"operation_summary":{"service_id":3977,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":2.00000000,"min_output_million_tokens_price":3.00000000,"max_price":0.0000,"max_input_million_tokens_price":2.00000000,"max_output_million_tokens_price":3.00000000,"free_operation_count":2,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":556,"account":0,"account_type":0,"type":"text2text","service_id":3977,"service_type":"serverless","model_id":24012,"vendor":"baidu-cloud","region":"prod-bce-default-ip","compute":"baidu-cloud/prod-bce-default-ip/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":2.00000000,"output_million_tokens_price":3.00000000,"input_million_tokens_price_average":2.00000000,"output_million_tokens_price_average":3.00000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"think","type":"boolean","control":"boolean","required":false,"description":"开启thinking","fixed_value":false},{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1766055091000,"updated_at":1786965473000,"vendor_info":{"id":0,"name":"baidu-cloud","label":"百度智能云","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-w36p9kcyq-3412a51e.png"},"region_info":{"name":"prod-bce-default-ip","label":"百度智能云默认","props":{}},"compute_info":{"id":"baidu-cloud/prod-bce-default-ip/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"prod-bce-default-ip","vendor_name":"baidu-cloud","type":"SERVERLESS_API","storage":0},"ids":[556],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216},{"id":557,"account":0,"account_type":0,"type":"text2text","service_id":3977,"service_type":"serverless","model_id":24012,"vendor":"baidu-cloud","region":"prod-bce-default-ip","compute":"baidu-cloud/prod-bce-default-ip/serverless_api/gpu/large/default","name":"文本对话（THINKING）","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":2.00000000,"output_million_tokens_price":3.00000000,"input_million_tokens_price_average":2.00000000,"output_million_tokens_price_average":3.00000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"think","type":"boolean","control":"boolean","required":false,"description":"开启thinking","fixed_value":true},{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1766055532000,"updated_at":1786965473000,"vendor_info":{"id":0,"name":"baidu-cloud","label":"百度智能云","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-w36p9kcyq-3412a51e.png"},"region_info":{"name":"prod-bce-default-ip","label":"百度智能云默认","props":{}},"compute_info":{"id":"baidu-cloud/prod-bce-default-ip/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"prod-bce-default-ip","vendor_name":"baidu-cloud","type":"SERVERLESS_API","storage":0},"ids":[557],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":3992,"ident":"HunyuanOCR","type":"serverless","vendor":"","region":"","compute":"","model":23784,"name":"HunyuanOCR","remark":"HunyuanOCR 是腾讯混元大模型体系下的通用光学字符识别（OCR）能力，基于多模态理解模型构建，具备高精度文本提取、版面还原与复杂场景识别能力。它支持多语言识别、复杂背景下文字提取，以及对表格、票据、证照等结构化信息的解析，适用于金融、政务、电商、教育等多类场景。","price":0.0000,"props":{"name":"HunyuanOCR","service_type":"serverless"},"inventories":0,"created_at":1764730868000,"updated_at":1764730939000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/news_cover_1-avatar-hunyuan.png","description_html":"HunyuanOCR 是腾讯混元大模型体系下的通用光学字符识别（OCR）能力，基于多模态理解模型构建，具备高精度文本提取、版面还原与复杂场景识别能力。它支持多语言识别、复杂背景下文字提取，以及对表格、票据、证照等结构化信息的解析，适用于金融、政务、电商、教育等多类场景。
HunyuanOCR 具有以下特点：
高精度文字识别：对自然场景、扫描文档、截图等多类图片具有强鲁棒性。
版面理解能力强：可自动识别段落、标题、表格等结构，支持复杂排版场景。
多语言支持：覆盖中英文及多语种混合识别。
格式解析能力：可识别并结构化提取票据、表格、合同等常见文档内容。
低延迟高性能：基于混元多模态模型优化，适合高并发业务接入。
HunyuanOCR 能够为企业提供灵活高效的图片文字数字化能力，是构建智能文档处理、内容审核、自动化流程等应用场景的核心组件。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1175,"name":"文档处理/OCR","slug":"document_processing_label","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-document.svg","en_name":"Document/OCR","color":"","description":"","href":"","created_at":1735902954339,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":23784,"name":"HunyuanOCR","namespace":"hf-models","mirror_namespace":"tencent","mirror_url":"https://huggingface.co/tencent/HunyuanOCR"},"total_calls":72,"operation_summary":{"service_id":3992,"unit_tag":0,"min_price":0.0200,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0200,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":520,"account":0,"account_type":0,"type":"image2text","service_id":3992,"service_type":"serverless","model_id":23784,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"OCR","description":"","path":"v1/images/ocr","protocol":"http","price":0.0200,"average_price":0.0200,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"form","name":"image","type":"file","accept":".jpg, .jpeg, .png, .gif, .webp","control":"file","samples":[{"label":"文字检测和识别","value":{"url":"https://gitee-ai.su.bcebos.com/samples/images/hyocr_demo_1.png"}},{"label":"文档解析","value":{"url":"https://gitee-ai.su.bcebos.com/samples/images/hyocr_demo_2.png"}},{"label":"信息抽取","value":{"url":"https://gitee-ai.su.bcebos.com/samples/images/hyocr_demo_3.png"}},{"label":"视觉问答","value":{"url":"https://gitee-ai.su.bcebos.com/samples/images/hyocr_demo_4.png"}}],"category":"main","required":true,"description":"图片，最大分辨率：2048x2048"},{"in":"form","name":"prompt","type":"string","control":"text","samples":[{"label":"文字检测和识别","value":"检测并识别图片中的文字，将文本坐标格式化输出。"},{"label":"文档解析","value":"提取文档图片中正文的所有信息用markdown 格式表示，其中页眉、页脚部分忽略，表格用html 格式表达，文档中公式用latex 格式表示，按照阅读顺序组织进行解析。"},{"label":"信息抽取","value":"提取图片中的：['单价', '上车时间', '发票号码', '省前缀', '总金额', '发票代码', '下车时间', '里程数'] 的字段内容，并按照 JSON 格式返回。"},{"label":"视觉问答","value":"What is the highest life expectancy at birth of male?"}],"category":"main","required":false,"description":"提示词"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1764730868000,"updated_at":1768563816000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[520],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":4284,"ident":"comfyui","type":"serverless","vendor":"","region":"","compute":"","model":0,"name":"ComfyUI","remark":"1","price":0.0000,"props":{"service_type":"serverless"},"inventories":0,"created_at":1762851549000,"updated_at":1762851549000,"sort_order":0,"status":1,"recommend":false,"cover":"","description_html":"Comfyui
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","total_calls":7,"operation_summary":{"service_id":4284,"unit_tag":0,"min_price":0.0050,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0300,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":4},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":528,"account":0,"account_type":0,"type":"comfyui_sampler","service_id":4284,"service_type":"serverless","model_id":0,"vendor":"metax","region":"prod-mx-gz-ip","compute":"metax/prod-metax-vms/serverless_api/gpu/xlarge/comfyui","name":"ComfyUI 采样器","description":"ComfyUI Sampler 采样器接口，用于图像生成采样","path":"v1/serverless-comfyui/sampler","protocol":"http","price":0.0300,"average_price":0.0300,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":1,"params":[],"status":"available","free_use":true,"api_format":"GITEE","created_at":1763023919000,"updated_at":1765278766000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-gz-ip","label":"广州","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-metax-vms/serverless_api/gpu/xlarge/comfyui","instance_size":"xlarge","instance_type":"comfyui","price_per_hour":20.00,"status":"unAvailable","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"comfyui","gpu_memory_gb":64,"region_name":"prod-metax-vms","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[528],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":529,"account":0,"account_type":0,"type":"comfyui_text_encode","service_id":4284,"service_type":"serverless","model_id":0,"vendor":"metax","region":"prod-mx-gz-ip","compute":"metax/prod-metax-vms/serverless_api/gpu/xlarge/comfyui","name":"ComfyUI 文本编码","description":"ComfyUI Text Encode 文本编码接口，用于文本特征提取","path":"v1/serverless-comfyui/text-encode","protocol":"http","price":0.0050,"average_price":0.0050,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":2,"params":[],"status":"available","free_use":true,"api_format":"GITEE","created_at":1763023919000,"updated_at":1765278767000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-gz-ip","label":"广州","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-metax-vms/serverless_api/gpu/xlarge/comfyui","instance_size":"xlarge","instance_type":"comfyui","price_per_hour":20.00,"status":"unAvailable","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"comfyui","gpu_memory_gb":64,"region_name":"prod-metax-vms","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[529],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":530,"account":0,"account_type":0,"type":"comfyui_vae_encode","service_id":4284,"service_type":"serverless","model_id":0,"vendor":"metax","region":"prod-mx-gz-ip","compute":"metax/prod-metax-vms/serverless_api/gpu/xlarge/comfyui","name":"ComfyUI VAE 编码","description":"ComfyUI VAE Encode 图像编码接口，用于图像到潜空间的转换","path":"v1/serverless-comfyui/vae/encode","protocol":"http","price":0.0050,"average_price":0.0050,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":3,"params":[],"status":"available","free_use":true,"api_format":"GITEE","created_at":1763023920000,"updated_at":1765278768000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-gz-ip","label":"广州","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-metax-vms/serverless_api/gpu/xlarge/comfyui","instance_size":"xlarge","instance_type":"comfyui","price_per_hour":20.00,"status":"unAvailable","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"comfyui","gpu_memory_gb":64,"region_name":"prod-metax-vms","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[530],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":531,"account":0,"account_type":0,"type":"comfyui_vae_decode","service_id":4284,"service_type":"serverless","model_id":0,"vendor":"metax","region":"prod-mx-gz-ip","compute":"metax/prod-metax-vms/serverless_api/gpu/xlarge/comfyui","name":"ComfyUI VAE 解码","description":"ComfyUI VAE Decode 图像解码接口，用于潜空间到图像的转换","path":"v1/serverless-comfyui/vae/decode","protocol":"http","price":0.0100,"average_price":0.0100,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":4,"params":[],"status":"available","free_use":true,"api_format":"GITEE","created_at":1763023920000,"updated_at":1765278769000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-gz-ip","label":"广州","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-metax-vms/serverless_api/gpu/xlarge/comfyui","instance_size":"xlarge","instance_type":"comfyui","price_per_hour":20.00,"status":"unAvailable","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"comfyui","gpu_memory_gb":64,"region_name":"prod-metax-vms","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[531],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":4309,"ident":"LongCat-Image","type":"serverless","vendor":"","region":"","compute":"","model":24054,"name":"LongCat-Image","remark":"LongCat-Image 是美团 LongCat 团队开源的轻量级高质量图像生成模型，支持中英文文本生成逼真图像，并具备优秀的中文文字渲染能力。","price":0.0000,"props":{"name":"LongCat-Image","service_type":"serverless"},"inventories":0,"created_at":1765187832000,"updated_at":1765187832000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-urqh01ecj-avatar-longcat.png","description_html":"LongCat-Image
LongCat-Image 是美团 LongCat 团队开源的高质量图像生成模型，具备 Text-to-Image（文本生成图片）能力，支持中英文 Prompt，能够根据自然语言描述生成逼真、风格化或创意类图像。模型在人物细节、背景纹理、光影表现上较为稳定，同时 具有优秀的中文字符渲染能力，可输出含中文文字的广告、包装、海报等视觉素材。
核心能力
文本生成图像（Text-to-Image）：根据中英文描述自动构建场景、人物、产品或插画内容；
高质量真实感输出：面部、纹理、光影一致性优异，可生成写实或卡通风格图像；
中文文字渲染能力强：在同类开源模型中表现领先，可稳定生成中文招牌、产品标签、文案、UI 字体等；
多风格支持：卡通、插画、写实、美术、海报等多种表现形式，可通过 Prompt 控制；
输入尺寸灵活：支持不同长宽比，适用于封面、海报、广告、电商素材等场景；
轻量高效推理：模型规模更精简，适合在 Serverless 场景低成本调用，也适用于移动端或网页 AI 创作工具。
典型应用场景
场景 | 说明 |
内容创作与素材生成 | 为短视频封面、公众号封面、社交媒体内容生成插画、角色、海报、配图等视觉素材 |
电商与产品展示 | 为商品设计、包装标签、广告、宣传图快速生成样图，并支持中文文字渲染 |
AI 绘画与角色概念设计 | 根据文本描述生成角色形象、设定草图、背景插画或故事分镜 |
营销与海报生成 | 可生成包含标题文案、广告牌、中文字体素材的视觉创意画面 |
设计工具与教育内容 | 可嵌入原型设计工具、图文编辑器、课程系统，实现即时图像生成体验 |
LongCat-Image-Edit
LongCat-Image-Edit 基于美团 LongCat 团队开源的 60 亿参数图像编辑模型构建，专注于自然语言驱动的图像重绘、内容修改与视觉替换。模型具备强语义理解能力，可在不破坏主体结构与整体风格的前提下，对局部或整图进行精准编辑，并支持参考图示例对齐，使生成结果在细节、纹理、比例与光影方面高度一致。
\uD83D\uDCCC 应用场景
应用场景 | 说明 | 典型效果 |
商品图修改与增强 | 支持根据文本描述替换背景、材质、光影或局部属性，适用于电商商品图内容提升 | 让商品更突出、更干净、更吸引 |
头像与人像优化 | 可替换五官、发型、妆容、服饰或背景，同时保持人物结构自然一致 | 社交头像、虚拟人形象、职场证件照 |
创意广告与海报生成 | 基于提示词进行全局构图调整或视觉风格转换，用于品牌营销与创意内容制作 | 快速输出不同风格的视觉素材 |
UI、插画与图标重绘 | 在保持主体布局的前提下进行扩展、风格化或角色替换 | 产品 UI 预览、品牌插画、App 图标 |
局部修复 / 涂抹去除 | 可对局部区域进行修补、抹除或细节填充，效果自然无破绽 | 去掉水印、杂物、人脸修补 |
HuatuoGPT-o1-7B是由FreedomIntelligence开发的医疗领域大型语言模型（LLM），专为高级医疗推理设计。该模型在提供最终回答之前，会生成复杂的思考过程，反映并完善其推理。HuatuoGPT-o1-7B支持中英文，能够处理复杂的医疗问题，并以'思考-回答'的格式输出结果，这对于提高医疗决策的透明度和可靠性至关重要。该模型基于Qwen2.5-7B，经过特殊训练以适应医疗领域的需求。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1150,"name":"32K","slug":"context_size_32k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-f8crosrdk-id-y4z8cy4z9-context_size.png","en_name":"32K","color":"blue","description":"上下文大小","href":"","created_at":1734693847615,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1210,"name":"行业大模型","slug":"Industry-specific Large Model","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-81573em6o-yuan.png","en_name":"Industry AI Models","color":"","description":"","href":"","created_at":1742979183278,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1214,"name":"医疗","slug":"medical_label","icon":"","en_name":"Medical","color":"","description":"","href":"","created_at":1745831902005,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1276,"name":"AI for Science","slug":"AI4S","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-58jwavsj1-flask-conical.svg","en_name":"AI for Science","color":"","description":"","href":"","created_at":1773294837459,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":24158,"name":"HuatuoGPT-o1-7B","namespace":"hf-models","mirror_namespace":"FreedomIntelligence","mirror_url":"https://huggingface.co/FreedomIntelligence/HuatuoGPT-o1-7B"},"total_calls":21,"operation_summary":{"service_id":4408,"unit_tag":0,"min_price":0.0000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0000,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":540,"account":0,"account_type":0,"type":"text2text","service_id":4408,"service_type":"serverless","model_id":24158,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1765338418000,"updated_at":1770658677000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[540],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":4423,"ident":"GLM-ASR","type":"serverless","vendor":"","region":"","compute":"","model":24163,"name":"GLM-ASR","remark":"GLM-ASR是一款超轻量级中文语音识别模型，提供低资源设备上的实时转写能力，适用于移动端、IoT 与本地离线识别场景。","price":0.0000,"props":{"name":"GLM-ASR-Nano-2512","service_type":"serverless"},"inventories":0,"created_at":1765354272000,"updated_at":1766128305000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-vs79e29q6-avatar-zhipuai.png","description_html":"","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":240,"name":"自动语音识别","slug":"automatic-speech-recognition","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-automatic-speech-recognition.svg","en_name":"Automatic Speech Recognition","color":"","description":"","href":"","created_at":1700443954790,"sort_order":0,"categories":[{"id":13,"parent_id":1,"name":"音频","slug":"audio","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/audio.svg","sort_order":13,"en_name":"Audio"},{"id":19,"parent_id":8,"name":"音频","slug":"dataset-audio","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-audio.svg","sort_order":19,"en_name":"Audio"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":24163,"name":"GLM-ASR-Nano-2512","namespace":"hf-models","mirror_namespace":"zai-org","mirror_url":"https://huggingface.co/zai-org/GLM-ASR-Nano-2512"},"total_calls":10,"operation_summary":{"service_id":4423,"unit_tag":0,"min_price":0.0000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0000,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":542,"account":0,"account_type":0,"type":"speech2text","service_id":4423,"service_type":"serverless","model_id":24163,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"语音识别","description":"","path":"v1/audio/transcriptions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"form","name":"file","type":"string","props":{"enableRecorder":false},"accept":".mp3, .wav","control":"file","category":"main","required":true,"description":"音频文件"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1765354327000,"updated_at":1783079286000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[542],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":4830,"ident":"AutoGLM-Phone-9B-Multilingual","type":"serverless","vendor":"","region":"","compute":"","model":24159,"name":"AutoGLM-Phone-9B-Multilingual","remark":"AutoGLM-Phone-9B-Multilingual 是一个基于 AutoGLM 的多语言手机智能助手大模型，能够通过视觉-语言理解手机屏幕内容并根据自然语言指令自动执行操作任务（如打开应用、浏览界面、点击等）。","price":0.0000,"props":{"name":"AutoGLM-Phone-9B-Multilingual","service_type":"serverless"},"inventories":0,"created_at":1765445001000,"updated_at":1766128301000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-vs79e29q6-avatar-zhipuai.png","description_html":"AutoGLM-Phone-9B-Multilingual
模型简介
AutoGLM-Phone-9B-Multilingual 是一款面向智能手机场景的多模态大模型，通过视觉理解（手机屏幕截图 + OCR）与多语言自然语言指令解析，实现对手机界面的自动操作与任务执行。该模型能够在「看懂手机界面」与「执行用户目标」之间建立智能推理链，适用于各类自动化、测试与智能助手产品。
模型核心特点
1. 多模态理解手机界面
同时处理 屏幕截图 + 文本区域（OCR）
能够识别图标、按钮、菜单、弹窗、列表内容
支持中文、英文等多语言界面理解
能跨设备、跨 APP 理解界面结构
2. 直接生成可执行的手机操作序列
模型不仅能回答“如何做”，还能输出可直接执行的操作步骤，包括：
点击（click）
滑动（swipe）
输入文本（input）
打开应用（launch）
页面导航（navigate）
这使得开发者可以用自然语言对模型下达命令，让模型自动完成任务。
3. 强大的任务规划能力
能够理解复杂目标，例如：
“打开微信并给张三发一条消息”
“查找抖音里的关注列表”
“设置手机为静音模式”
模型会根据界面状态动态规划最优路径。
4. 多语言支持（Multilingual）
不仅能理解各语言指令，还能对多语言 UI 做精准识别和指令映射，适合跨国家产品。
5. 极强的可泛化能力
AutoGLM 体系无需对每个 App 单独训练即可理解其界面逻辑，适用于：
不同品牌手机（Android 主流厂商）
不同主题/不同语言
各类 App（社交、电商、系统设置、工具类等）
使用场景
场景类别 | 说明与价值 |
移动 App 自动化测试 | 自动执行点击、滑动、输入等动作，用于 UI 测试、回归测试、冒烟测试，降低人工成本，提高测试效率。 |
智能手机助手 / 语音助手 | 用户说一句话即完成复杂操作，如“帮我打开支付宝扫一扫”“把手机调成飞行模式”，提升智能助手能力。 |
无障碍辅助（Accessibility） | 通过语言即可让模型操作手机界面，为视障人士提供更自然的交互能力。 |
企业自动化流程（RPA on Mobile） | 自动化处理移动端 ERP、OA、审批、营销工具操作，减少手工反复操作。 |
App 内容采集 / 分析机器人 | 自动进入页面、滚动、提取内容，用于采集数据或执行自动化流程。 |
移动端演示 / 训练代理 | 自动执行一套标准流程，如教学引导、录制演示视频、功能展示。 |
客服辅助定位问题 | 客服给定指令，如“告诉我用户应该怎么找到该功能按钮”，模型生成可执行步骤并标注界面区域。 |
InfiniteTalk 是 MeiGen-AI 推出的新一代音频驱动视频生成模型，专注于将语音内容转化为自然、连贯、可无限时长扩展的数字人讲述视频。模型基于稀疏帧视频建模与音视频跨模态对齐技术，在保证高质量口型同步的同时，大幅降低长视频生成过程中的计算与漂移问题。
在使用层面，InfiniteTalk 以 音频作为唯一时间轴驱动源。生成视频的整体时长 严格由上传的音频时长决定：音频多长，生成的视频即对应多长，不存在固定时长限制或强制截断。模型会逐帧对齐音频信号，在整个音频播放期间持续驱动口型、表情和头部动作，实现从音频起始到结束的完整视频生成。
在核心能力上，InfiniteTalk 具备以下特点：
高精度口型同步
能够根据输入语音精确生成与发音高度匹配的嘴部动作，适配不同语言、语速和说话风格，显著降低“对不上嘴”的违和感。自然的头部与上半身动作建模
除了嘴部，模型还能生成符合语义与语气的头部摆动、微表情和上半身姿态变化，使人物表现更加生动真实。无限时长视频生成能力
通过稀疏帧和时序一致性设计，模型可稳定生成超长视频内容，避免传统方法中常见的画面抖动、表情崩坏或人物“漂移”问题。输入形式灵活
支持以单张图片或短视频作为人物形象基础，结合任意长度的音频，快速生成完整的讲述视频，适配多种内容生产流程。风格与人物泛化能力强
对不同人脸风格、年龄、性别和画面风格具备良好的泛化能力，适合构建多样化数字人角色库。
使用方式概览
单人模式（Single）
用途：让单个角色（或者单张人物图像）“说话”或“唱歌”。
过程：
准备一个人物图片或已有视频。
上传音频（或先用 TTS 生成音频）。
系统根据音频驱动模型生成视频输出。
下载生成的 MP4 结果。
特点：
适合独白、讲解类内容。
生成自然口型、头部和表情动作。
通常默认 480p 或 720p 输出。
多人模式（Multi）
用途：同时处理 两个或更多角色的对话视频。
过程：
同样上传一张或多张人物图像。
每个角色对应一条独立的音频轨道。
系统自动将音频和图像对应起来，根据时间轴合成视频。
输出带有多角色对话、自然动作的视频。
特点：
支持多角色轮流发言、对话流程自然。
忠实还原不同音频对应的唇型、表情、互动感觉。
适用于访谈、对话、双人表演等场景。
应用场景说明
凭借其高稳定性和长时生成能力，InfiniteTalk 在多个实际场景中具备较高落地价值：
1. 数字人讲解与虚拟主播
适用于新闻播报、财经解读、产品介绍、企业宣传等场景，低成本生成可长时间连续输出内容的虚拟主播视频。
2. 在线教育与知识科普
将课程音频、讲解文稿快速转化为“会说话的老师”视频，用于录播课程、培训内容和科普视频制作。
3. AI 客服与虚拟助手
结合语音合成与对话系统，生成具备自然表情和肢体语言的虚拟客服或数字助手，提升交互真实感。
4. 内容创作者与自媒体生产
帮助个人创作者或团队快速批量生成讲述类视频，如解说、故事讲述、知识分享等，大幅降低视频制作门槛。
5. 数字人 IP 与虚拟形象运营
用于打造可长期运营的数字人 IP，实现持续输出视频内容，适合虚拟偶像、品牌代言人等方向。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1191,"name":"视频生成","slug":"video-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text_to_video.svg","en_name":"Video Generation","color":"","description":"","href":"","created_at":1739973857876,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1224,"name":"数字人","slug":"Digital_Human","icon":"","en_name":"Digital Human","color":"","description":"","href":"","created_at":1750325639154,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":23745,"name":"InfiniteTalk","namespace":"hf-models","mirror_namespace":"MeiGen-AI","mirror_url":"https://huggingface.co/MeiGen-AI/InfiniteTalk"},"total_calls":21,"operation_summary":{"service_id":4982,"unit_tag":0,"min_price":0.5000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.5000,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":550,"account":0,"account_type":0,"type":"image2video","service_id":4982,"service_type":"serverless","model_id":23745,"vendor":"gitee","region":"prod-gitee-urc-ip","compute":"gitee/prod-gitee-urc-ip/serverless_api/gpu/large/default","name":"图生视频","description":"","path":"v1/async/videos/image-to-video","protocol":"http","price":0.5000,"average_price":0.5000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"form","name":"image","type":"file","label":"参考图","control":"file","visible":false,"category":"main","required":false,"description":"用于引导生成预期图像的描述文字。"},{"in":"form","name":"cond_video","type":"string","accept":"image/*, video/*","control":"file","samples":[{"label":"示例一","value":{"url":"https://gitee-ai.su.bcebos.com/samples/images/woman-singing.png","name":"man-woman-in-car.png"}}],"category":"main","required":true,"description":"生成视频所需的图片"},{"in":"form","name":"cond_audio","type":"array","items":{"type":"file","control":"file"},"props":{"fileField":{"max_length":2},"textField":false,"imageField":false},"accept":"audio/*","control":"multimodal","example":["/path/to/audio.wav"],"samples":[{"label":"示例一","value":[{"type":"files","value":["https://gitee-ai.su.bcebos.com/samples/images/examples_single.wav"]}]}],"category":"main","required":true,"max_length":2,"description":"生成视频所需的音频文件，最大支持15秒的音频"},{"in":"form","name":"prompt","type":"string","control":"text","samples":[{"label":"示例一","value":"A woman is passionately singing into a professional microphone in a recording studio. She wears large black headphones and a dark cardigan over a gray top. Her long, wavy brown hair frames her face as she looks slightly upwards, her mouth open mid-song. The studio is equipped with various audio equipment, including a mixing console and a keyboard, with soundproofing panels on the walls. The lighting is warm and focused on her, creating a professional and intimate atmosphere. A close-up shot captures her expressive performance."}],"category":"main","required":true,"description":""},{"in":"form","name":"num_inference_steps","step":1,"type":"integer","control":"slider","maximum":10,"minimum":0,"required":false,"description":"生成视频时模型执行迭代的次数。数值越大，生成的质量会更高","default_value":4},{"in":"form","name":"motion_frame","step":1,"type":"integer","control":"slider","maximum":15,"minimum":1,"required":false,"description":"用于长视频生成模式的驱动帧长度。","default_value":9},{"in":"form","name":"seed","type":"integer","control":"number","required":false,"description":"用于确保随机过程的结果可复现，从而保证实验的一致性。"},{"in":"body","name":"size","type":"string","control":"select","options":[{"label":"infinitetalk-480","value":"infinitetalk-480"},{"label":"infinitetalk-720","value":"infinitetalk-720"}],"required":false,"description":"尺寸","default_value":"infinitetalk-480"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1765787406000,"updated_at":1783076576000,"vendor_info":{"id":0,"name":"gitee","label":"Gitee","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8n73csd10-image 5.svg"},"region_info":{"name":"prod-gitee-urc-ip","label":"乌鲁木齐","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"gitee/prod-gitee-urc-ip/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"NVIDIA系列","gpu_memory_gb":24,"gpu_performance_rate":"full","region_name":"prod-gitee-urc-ip","vendor_name":"gitee","type":"SERVERLESS_API","storage":0},"ids":[550],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":5011,"ident":"GLM-TTS","type":"serverless","vendor":"","region":"","compute":"","model":24216,"name":"GLM-TTS","remark":"GLM-TTS 是一个支持零样本克隆音色、情感可控和实时推理的高质量文本到语音（TTS）合成系统。","price":0.0000,"props":{"name":"GLM-TTS","service_type":"serverless"},"inventories":0,"created_at":1765866139000,"updated_at":1766128269000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-vs79e29q6-avatar-zhipuai.png","description_html":"GLM-TTS 模型详细介绍
一、模型特点
GLM-TTS 是智谱推出的高质量文本转语音（TTS）模型，具备优秀的音色泛化能力与自然语音生成效果，特别适合通过「参考音频 + 文本」的方式实现零样本音色复用与稳定语音合成。
模型核心特点包括：
基于参考音频的音色保持能力
通过上传一段参考语音，模型可在不训练、不微调的情况下，生成音色高度一致的语音内容，适合快速复刻说话人风格。文本与语音语义对齐准确
对输入文本理解准确，断句自然，语调与语义匹配度高，适用于中长文本语音生成。生成语音自然、清晰
语音发音清楚，韵律自然，适合正式播报、讲述类场景。参数简洁，易于工程集成
仅需提供参考音频 URL 与对应文本即可生成目标语音，降低接入与使用复杂度，适合标准化 API 服务输出。
二、参数使用方式（API 说明）
我们提供的 GLM-TTS API 目前包含以下三个核心参数：
1. prompt_audio_url
含义：参考音频地址
作用：用于指定目标生成语音的音色和说话风格
说明：
音频内容建议为单一说话人
音质清晰、无明显背景噪声效果更佳
音频时长可较短，用于提供音色参考即可
2. prompt_text
含义：参考音频对应的文本内容
作用：帮助模型对齐参考音频与文字语义，提高音色和发音一致性
说明：
文本内容应与
prompt_audio_url中的语音语义一致用于指导模型学习说话人的发音方式与语调特征
3. output_format
含义：生成音频的输出格式
常见取值：
wav、mp3作用：控制最终返回的音频文件格式，便于不同业务场景直接使用
通过「参考音频 + 对应文本」的方式，API 可在保持音色一致性的前提下，生成新的语音内容，适合零样本音色复用场景。
三、应用场景
在当前参数体系下，GLM-TTS 特别适合以下应用方向：
1. 数字人 / 虚拟角色语音
数字人视频配音
虚拟主播、虚拟讲解员
AI 讲述类视频内容生成
2. 内容语音化生产
有声内容批量生成
课程、培训材料语音化
文档、资讯自动播报
3. 企业与产品语音输出
固定音色的系统语音提示
企业内部系统语音播报
产品内统一风格语音输出
4. 音色复刻与个性化语音
快速复刻指定说话人音色
个性化语音内容生成
无需训练的音色迁移应用
一、模型特点
1. 新一代高一致性图像生成模型
FLUX.2-dev 在构图稳定性与画面一致性方面表现突出，能够在多次生成中保持人物结构、风格特征和整体视觉逻辑，有效降低“跑偏”和细节崩坏问题。
2. 优秀的文本语义对齐能力
模型对复杂 Prompt、组合指令和抽象概念具备较强理解能力，能够准确将文字描述映射为视觉元素，适合需要精确控制生成结果的应用场景。
3. 面向真实生产环境优化
FLUX.2-dev 在生成质量与推理效率之间取得平衡，适合在实际业务中进行高频调用，而不仅限于实验或展示用途。
4. 支持生成与再创作类任务
模型可用于从零生成图像，也适用于在既有视觉基础上进行风格调整、画面优化和内容重构，满足多阶段创作流程需求。
二、使用场景
1. 商业级视觉内容生产
广告素材、品牌视觉图生成
电商商品展示图、营销海报
高质量社交媒体内容创作
2. 设计与创意辅助工具
设计师灵感生成与方案探索
概念图、草图到成图的快速迭代
多风格、多方案并行生成
3. 图像生成与编辑平台
文生图 / 图生图能力模块
AI 创作平台的核心生成模型
多模型体系中的高质量图像选项
GLM-4.7 是由 智谱 AI 发布的最新一代大型语言模型（358B），继承并扩展了 GLM-4 系列模型的核心能力，是面向高复杂任务（尤其是编码、推理与代理任务）的旗舰级开源模型。
\uD83D\uDCCC 核心定位
旗舰级通用模型：GLM-4.7 在 GLM-4.6 的基础上进一步提升了多方面性能，尤其是在代码生成、复杂推理、工具调用和智能代理场景中表现显著更强。
适合高级开发与集成：可用于构建智能编程助手、自动化工具链、长上下文理解系统、复杂角色对话及创意生成等应用。
\uD83D\uDE80 主要能力与特性
✔ 编码能力提升
相比上一代 GLM-4.6，GLM-4.7 在多语言编码任务（如 SWE-bench、Terminal Bench）上显著提高性能，支持更强的 agentic 编程与工具交互能力。
✔ 综合推理更强
在数学推理与高级逻辑任务中也有明显提升，如在 HLE（Humanity’s Last Exam）等复杂评测中成绩提升显著。
✔ 工具使用与代理能力
强化了模型对外部工具调用（Tool Using）与浏览等增强模块的结合能力，使其在具有动态任务分解与执行场景中更加稳定和有效。
✔ 更好的内容生成质量
不仅限于技术任务，GLM-4.7 在一般对话、创意写作、角色扮演等生成质量上也有提升，使交互体验更流畅。
\uD83D\uDD27 技术特点
超大 Context 支持：可以处理超长上下文，对大型文档、长会话等场景友好（具体上下文长度依配置而定）。
思维控制机制：引入 “Interleaved Thinking（交错思考）”、“Preserved Thinking（保持思考）” 与 “Turn-level Thinking（回合级思考）” 等机制，提升复杂任务中的一致性和控制能力。
开放许可：模型及权重在 Hugging Face 采用 MIT 许可，可用于商业/研究项目。
\uD83D\uDCE6 使用方式
可以通过主流推理框架（如 transformers、vLLM、SGLang）加载模型进行本地推理或部署，同时支持标准对话调用和高级工具调用格式。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":33,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1262,"name":"200K","slug":"context_size_200k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-v6ul4uqrf-id-9zvdx4rar-id-zd1snxypr-context_size.png","en_name":"200K","color":"blue","description":"上下文大小","href":"","created_at":1759218439298,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1287,"name":"梯度计费","slug":"tiered_pricing","icon":"","en_name":"Tiered Pricing","color":"","description":"","href":"","created_at":1787913527124,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":24354,"name":"GLM-4.7","namespace":"hf-models","mirror_namespace":"zai-org","mirror_url":"https://huggingface.co/zai-org/GLM-4.7"},"total_calls":1211,"operation_summary":{"service_id":5360,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":3.50000000,"min_output_million_tokens_price":14.00000000,"max_price":0.0000,"max_input_million_tokens_price":3.50000000,"max_output_million_tokens_price":14.00000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":761,"account":0,"account_type":0,"type":"text2text","service_id":5360,"service_type":"serverless","model_id":24354,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":3.50000000,"output_million_tokens_price":14.00000000,"input_million_tokens_price_average":3.50000000,"output_million_tokens_price_average":14.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"tiered_prompt_cache_tokens","tiers":[{"average":{"input_million_tokens_price":3,"output_million_tokens_price":14,"cache_read_million_tokens_price":0.6},"specified":{"input_million_tokens_price":3,"output_million_tokens_price":14,"cache_read_million_tokens_price":0.6},"max_input_tokens":32000,"min_input_tokens":0},{"average":{"input_million_tokens_price":4,"output_million_tokens_price":16,"cache_read_million_tokens_price":0.8},"specified":{"input_million_tokens_price":4,"output_million_tokens_price":16,"cache_read_million_tokens_price":0.8},"min_input_tokens":32000}],"version":1,"calculation_mode":"whole_order"},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1780385189000,"updated_at":1787914967000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[761],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":5486,"ident":"MiniMax-M2.1","type":"serverless","vendor":"","region":"","compute":"","model":24420,"name":"MiniMax-M2.1","remark":"MiniMax-M2.1 是 MiniMax 系列的最新版本，专为多语言编程和真实世界复杂任务优化。作为一款 AI 原生模型，MiniMax-M2.1 在模型性能、智能体框架支持以及多场景适配方面实现了显著提升，旨在帮助企业和个人更快地找到 AI 原生的工作与生活方式。","price":0.0000,"props":{"name":"MiniMax-M2.1","service_type":"serverless"},"inventories":0,"created_at":1766567470000,"updated_at":1787316656000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-uvnkgppw9-id-3s2ivbvef-id-p30an22k1-dUd-LsZEX0H_d4qefO_g6.webp","description_html":"MiniMax-M2.1 是 MiniMax 系列的最新版本，专为多语言编程和真实世界复杂任务优化。作为一款 AI 原生模型，MiniMax-M2.1 在模型性能、智能体框架支持以及多场景适配方面实现了显著提升，旨在帮助企业和个人更快地找到 AI 原生的工作与生活方式。
在 MiniMax-M2 中，我们主要解决了模型成本和可访问性的问题。而在 MiniMax-M2.1 中，我们专注于提升模型在真实世界复杂任务中的表现，特别是在多编程语言和办公场景中的可用性，达到了业内领先水平。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":32,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1262,"name":"200K","slug":"context_size_200k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-v6ul4uqrf-id-9zvdx4rar-id-zd1snxypr-context_size.png","en_name":"200K","color":"blue","description":"上下文大小","href":"","created_at":1759218439298,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":24420,"name":"MiniMax-M2.1","namespace":"hf-models","mirror_namespace":"MiniMaxAI","mirror_url":"https://huggingface.co/MiniMaxAI/MiniMax-M2.1"},"total_calls":333,"operation_summary":{"service_id":5486,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":2.10000000,"min_output_million_tokens_price":8.40000000,"max_price":0.0000,"max_input_million_tokens_price":2.10000000,"max_output_million_tokens_price":8.40000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":702,"account":0,"account_type":0,"type":"text2text","service_id":5486,"service_type":"serverless","model_id":24420,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":2.10000000,"output_million_tokens_price":8.40000000,"input_million_tokens_price_average":2.10000000,"output_million_tokens_price_average":8.40000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":0.21,"cache_write_million_tokens_price":2.1},"version":1,"specified":{"cache_read_million_tokens_price":0.21,"cache_write_million_tokens_price":2.1}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1776840478000,"updated_at":1787316643000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[702],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":5621,"ident":"Qwen-Image-Edit-2511","type":"serverless","vendor":"","region":"","compute":"","model":24370,"name":"Qwen-Image-Edit-2511","remark":"Qwen-Image-Edit-2511 是阿里通义团队在 Qwen-Image-Edit 基础上升级的开源图像编辑模型，主打提升编辑一致性、集成常用 LoRA 能力并增强人物/工业设计和几何推理表现，实现更稳定精确的 AI 图像编辑效果。","price":0.0000,"props":{"name":"Qwen-Image-Edit-2511","service_type":"serverless"},"inventories":0,"created_at":1766744767000,"updated_at":1769215256000,"sort_order":0,"status":1,"recommend":true,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"模型简介
Qwen-Image-Edit-2511 是通义千问（Qwen）推出的高质量指令驱动图像编辑模型，面向真实业务场景优化，支持通过自然语言对现有图像进行精准、可控、结构保持的编辑。
模型在人物一致性、局部修改稳定性、复杂场景理解以及多轮编辑可复用性方面进行了重点增强，适合通过 API 集成到各类图像生产与处理系统中。
核心能力
1. 指令驱动的图像编辑（Instruction-based Editing）
使用自然语言描述修改需求，无需复杂参数配置
支持中文 / 英文指令
能准确理解「改什么」「不改什么」
示例
“把背景换成夜晚城市，但人物保持不变”
“把衣服颜色改成深蓝色，材质保持不变”
2. 局部编辑与结构保持
支持对指定区域或目标进行修改
最大限度保留原图结构、构图和细节
避免常见的整体风格漂移、人物变形问题
适用于：
局部重绘（Inpainting）
局部替换 / 增强
细节修复（去瑕疵、补全缺失区域）
3. 人物与主体一致性增强
对人脸、姿态、服装结构保持能力强
多次编辑同一张图时，主体不易“跑样”
适合证件照、模特图、IP 角色、品牌人物等场景
4. 风格与视觉调整
支持整体或局部风格修改：
插画 / 写实 / 工业设计风格
明亮 / 暗色 / 冷暖色调
在不破坏原有内容的前提下完成风格统一
5. 工业设计与 UI/产品图理解
对产品结构、几何关系、材质理解较好
适合修改：
产品外观颜色、材质
UI 素材中的图形元素
设计稿的局部调整
典型应用场景
1. 电商与营销内容生成
商品图背景替换
模特服装颜色 / 款式微调
多风格营销素材快速生成
2. 设计工具与创作平台
在线修图 / AI 设计助手
设计师辅助修改方案
多轮编辑、反复微调
3. 产品与工业设计
产品外观方案快速验证
材质、颜色、细节调整
概念图到展示图的迭代
4. 内容审核与自动化处理
自动修复不合规或瑕疵区域
批量图像统一风格
AI 驱动的图片处理流水线
模型特点（Fun-ASR-Nano-2512）
高精度语音识别
基于大规模真实语音数据训练，在通用语音及行业场景下均具备稳定、准确的转写能力。多语言与多口音支持
支持多种主流语言，并针对中文普通话及多地区口音、方言进行了优化。低延迟与实时转写能力
支持实时或近实时语音转文字，适合对时延敏感的在线应用。复杂音频场景适应性强
在噪声、远场、会议、音乐背景等复杂音频环境下仍保持较好的识别效果。轻量化与易集成
模型规模相对精简，推理效率高，适合 Serverless API、在线服务和批量处理场景。
使用场景
实时字幕与会议记录
用于视频会议、直播、课堂等场景的实时语音转写与自动纪要生成。语音助手与智能客服
支持多语言语音输入，适合构建客服系统、语音交互应用。音视频内容转写
对录音、播客、视频内容进行批量语音转文字处理。行业语音分析
适用于教育、金融、政务等领域的语音数据结构化与文本分析。
SAM 3（Segment Anything Model 3） 是新一代通用视觉分割模型，支持对图像中的任意目标进行高精度、像素级分割。模型无需固定类别训练，可通过视觉提示或文本语义对目标进行分割，适用于开放场景下的多目标识别与精细切割。
我们基于 SAM 3 封装并提供标准化 API 服务，支持高并发、低延迟调用，适合集成到各类视觉应用与业务系统中。
接口输出说明（Mask 编码）
为提升网络传输效率，图像分割接口返回的 mask 默认采用 COCO RLE（Run-Length Encoding）编码格式。
为什么使用 RLE？
相比 base64 bitmap，体积更小
非常适合大尺寸、高精度分割结果
是计算机视觉领域的通用标准（COCO、Detectron、SAM 等）
Mask 解码方式说明（前端 / 客户端）
接口返回的 mask 字段为 COCO RLE counts，需结合图像宽高进行解码，解码后可得到 H × W 的二维 0/1 mask 矩阵。
RLE 解码示例（JavaScript）
/**\n * 解码 COCO RLE 格式的 mask\n * @param rleString - RLE 编码字符串\n * @param height - 图像高度\n * @param width - 图像宽度\n * @returns 二维数组 [height][width]，1 表示前景，0 表示背景\n */\nexport function rleDecode(\n rleString: string,\n height: number,\n width: number,\n): number[][] {\n rleString = atob(rleString);\n // 解码 RLE counts\n const counts: number[] = [];\n let p = 0;\n while (p < rleString.length) {\n let x = 0;\n let k = 0;\n let more = true;\n while (more) {\n const c = rleString.charCodeAt(p) - 48;\n x |= (c & 0x1f) << (5 * k);\n more = (c & 0x20) !== 0;\n p++;\n k++;\n if (!more && (c & 0x10) !== 0) {\n x |= -1 << (5 * k);\n }\n }\n if (counts.length > 2) {\n x += counts[counts.length - 2];\n }\n counts.push(x);\n }\n // 从 RLE counts 创建扁平数组（列优先顺序）\n const total = height * width;\n const mask = new Uint8Array(total);\n let pos = 0;\n let val = 0;\n for (const count of counts) {\n for (let i = 0; i < count && pos < total; i++) {\n mask[pos] = val;\n pos++;\n }\n val = 1 - val; // 切换 0 和 1\n }\n // 将扁平数组转换为二维数组（列优先到行优先）\n const result: number[][] = Array.from({ length: height }, () =>\n new Array(width).fill(0),\n );\n // COCO RLE 使用列优先（Fortran）顺序存储\n let idx = 0;\n for (let x = 0; x < width; x++) {\n for (let y = 0; y < height; y++) {\n if (idx < total) {\n result[y][x] = mask[idx];\n idx++;\n }\n }\n }\n return result;\n}解码后的 mask[y][x]：
1表示该像素属于分割目标0表示背景
模型特点
1. 通用分割（Segment Anything）
不依赖预定义类别
可对任意物体、区域或前景进行分割
适用于复杂背景与未知目标场景
2. 像素级高精度 Mask 输出
输出精确到像素级的分割结果
支持返回 Mask（RLE 编码） 与 Bounding Box
适合对精度要求高的业务（如抠图、医学影像）
3. 多目标与复杂场景支持
单张图片可同时分割多个实例
支持目标重叠、小目标、边界复杂目标
稳定应对高分辨率图像
4. 工业级 API 设计
Serverless 架构，按调用量计费
接口稳定，易于前后端及算法系统集成
结果格式统一，便于二次处理与可视化
适用场景
1. 智能抠图与图像编辑
人像 / 商品 / 前景自动抠图
AI 设计工具、图片编辑器
2. 目标检测与视觉分析
自动标注训练数据
辅助目标检测、实例分割任务
3. 电商与内容审核
商品主体识别
背景剔除、商品区域提取
4. 医学影像与遥感分析
器官、病灶、地物区域分割
高精度像素级区域分析
5. 视频与多模态应用（扩展）
目标区域追踪（结合视频帧）
多模态 AI 系统的视觉输入模块
CosyVoice 3.0 是一款高质量、低延迟的多语言文本转语音（TTS）模型，支持中文、英文、日语、韩语等多种语言及多种中文方言，具备零样本声音克隆能力，仅需少量参考音频即可复刻说话人音色；模型支持流式语音生成，响应快、自然度高，并通过 instruct_text、speed、seed 等参数实现对语气、情绪、语速和生成一致性的精细控制。
使用场景
AI 语音助手 / 客服系统：自然、亲切的多语言语音交互，支持实时响应
内容播报与 AI 主播：新闻、公告、有声内容自动生成
教育与培训场景：课程讲解、语言学习、多语种朗读
跨语言配音与本地化：同一声音输出多语言内容，提升一致性
智能硬件与应用内语音：适配低延迟、高稳定性的在线 TTS 需求
VajraV1 模型详细介绍
VajraV1 是一套基于现代 YOLO 架构的高性能视觉理解模型，专注于实时计算机视觉任务，具备良好的准确性与推理效率平衡。模型支持多种核心视觉能力，可广泛应用于安防监控、工业检测、智慧零售、智能交通、AI 互动等场景。
我们基于 VajraV1 对外提供了 目标检测、图像分割、人体关键点检测 三类标准化 API 接口，便于快速集成和规模化部署。
一、目标检测
能力说明
目标检测接口用于在图片中自动识别并定位多个目标对象，返回每个目标的类别、置信度以及对应的边界框（Bounding Box）。
该能力可在单张图片中同时检测多类目标，适用于实时或批量场景。
典型应用场景
安防与监控：人、车辆、异常物体检测
智能交通：车辆识别、行人检测、道路目标分析
智慧零售：客流统计、商品识别、货架分析
工业视觉：零部件检测、缺陷目标识别
内容审核：特定物体或场景自动识别
二、图像分割
能力说明
图像分割接口可对图片中的目标进行像素级识别，输出精确的目标轮廓区域（Mask），而不仅仅是矩形框。
相比目标检测，分割能力能够提供更精细的空间信息。
接口输出说明（Mask 编码）
为提升网络传输效率，图像分割接口返回的 mask 默认采用 COCO RLE（Run-Length Encoding）编码格式。
为什么使用 RLE？
相比 base64 bitmap，体积更小
非常适合大尺寸、高精度分割结果
是计算机视觉领域的通用标准（COCO、Detectron、SAM 等）
Mask 解码方式说明（前端 / 客户端）
接口返回的 mask 字段为 COCO RLE counts，需结合图像宽高进行解码，解码后可得到 H × W 的二维 0/1 mask 矩阵。
RLE 解码示例（JavaScript）
/**\n * 解码 COCO RLE 格式的 mask\n * @param rleString - RLE 编码字符串\n * @param height - 图像高度\n * @param width - 图像宽度\n * @returns 二维数组 [height][width]，1 表示前景，0 表示背景\n */\nexport function rleDecode(\n rleString: string,\n height: number,\n width: number,\n): number[][] {\n rleString = atob(rleString);\n // 解码 RLE counts\n const counts: number[] = [];\n let p = 0;\n while (p < rleString.length) {\n let x = 0;\n let k = 0;\n let more = true;\n while (more) {\n const c = rleString.charCodeAt(p) - 48;\n x |= (c & 0x1f) << (5 * k);\n more = (c & 0x20) !== 0;\n p++;\n k++;\n if (!more && (c & 0x10) !== 0) {\n x |= -1 << (5 * k);\n }\n }\n if (counts.length > 2) {\n x += counts[counts.length - 2];\n }\n counts.push(x);\n }\n // 从 RLE counts 创建扁平数组（列优先顺序）\n const total = height * width;\n const mask = new Uint8Array(total);\n let pos = 0;\n let val = 0;\n for (const count of counts) {\n for (let i = 0; i < count && pos < total; i++) {\n mask[pos] = val;\n pos++;\n }\n val = 1 - val; // 切换 0 和 1\n }\n // 将扁平数组转换为二维数组（列优先到行优先）\n const result: number[][] = Array.from({ length: height }, () =>\n new Array(width).fill(0),\n );\n // COCO RLE 使用列优先（Fortran）顺序存储\n let idx = 0;\n for (let x = 0; x < width; x++) {\n for (let y = 0; y < height; y++) {\n if (idx < total) {\n result[y][x] = mask[idx];\n idx++;\n }\n }\n }\n return result;\n}解码后的 mask[y][x]：
1表示该像素属于分割目标0表示背景
典型应用场景
抠图与图像编辑：人物/物体精确抠图
自动驾驶与机器人：道路、障碍物、区域分割
工业检测：表面缺陷区域识别、精细质量检测
医学影像：组织、器官区域辅助分析
电商与内容生成：商品背景分离、素材处理
三、人体关键点检测
能力说明
人体关键点检测接口可自动识别图片中一个或多个人体，并返回人体各关键点（如头部、肩膀、手臂、腿部等）的坐标信息，用于描述人体姿态和动作结构。
典型应用场景
运动与健身：动作识别、姿态评估、训练辅助
智能安防：异常行为分析、跌倒检测
人机交互：体感交互、动作控制
娱乐与内容创作：动作捕捉、虚拟形象驱动
行为分析：人群行为研究、姿态统计分析
模型介绍
HY-MT1.5-7B 是腾讯混元（Hunyuan）团队推出并开源的 高性能多语言机器翻译模型，专为真实生产环境下的文本翻译任务设计。该模型基于 7B 参数规模训练，在多语言互译质量、专业领域适配能力和输出稳定性方面表现突出，适合通过 API 形式进行规模化调用和集成。
模型特点
1. 高质量多语言翻译能力
HY-MT1.5-7B 支持 33 种主流语言之间的双向翻译，在中英、英中以及多语对翻译任务上表现尤为稳定，能够准确理解上下文语义，减少直译、生硬翻译等问题，输出更加自然、符合目标语言表达习惯。
2. 面向真实业务场景优化
模型在训练和设计上充分考虑了 企业级翻译需求，在以下方面表现优异：
保留原文格式（段落、列表、标点、数字）
对专业术语、专有名词翻译更稳定
对长文本和多段落内容具备良好一致性
非常适合直接接入业务系统，而非仅用于实验或研究。
3. 上下文理解与一致性强
相比传统逐句翻译模型，HY-MT1.5-7B 对上下文有更强的整体理解能力，在文档级翻译中能够保持 术语前后一致、语义连贯，降低人工二次校对成本。
4. 可控性与工程友好性
模型适合以 API 形式部署和调用，便于进行：
批量翻译
流水线集成
与内容审核、搜索、内容生成等系统联动
在性能、稳定性和可维护性之间取得了良好平衡。
使用场景
\uD83C\uDF10 多语言内容翻译
网站、应用、产品文档的多语言本地化
新闻、资讯、社区内容的自动翻译
跨语言内容分发与国际化运营
\uD83E\uDDFE 企业文档与专业文本翻译
技术文档、API 文档、使用手册
商务材料、合同、报告（非法律最终稿）
行业术语较多的专业内容翻译
\uD83D\uDCAC 跨语言交流与客服
多语言客服消息实时翻译
工单、用户反馈的语言统一处理
跨语言沟通辅助系统
\uD83D\uDD0D 搜索与数据处理
多语言内容统一翻译后再进行检索、分析
作为 NLP 流水线中的翻译前置模块
Sinong1.0-32B 是由 NAULLM 团队发布的 320 亿参数开源农业大语言模型，围绕农业生产、农业科技与农业管理场景进行专项优化，重点提升农业专业知识理解、行业问答、场景推理与指令跟随能力，适合作为农业智能应用的核心模型能力底座。
模型核心特点
\uD83C\uDF3E 农业领域深度优化
面向农业语境进行专项训练与指令适配
覆盖作物种植、病虫害、防灾减灾、农资使用、农业政策等高频农业知识
对农业术语、生产流程与管理逻辑具备良好理解能力
\uD83E\uDDE0 32B 参数规模，支持复杂农业问题分析
适合多条件、多约束的农业决策类问题
能处理长文本农业资料（技术手册、政策文件、科研报告）
在农业知识整合、步骤推理与综合分析方面表现稳定
\uD83D\uDDE3 强中文能力，贴近农业实际表达
适配农业从业者、基层管理人员、技术推广人员的语言习惯
支持口语化咨询与专业化书面输出
可用于「专家式回答」与「通俗化解释」双模式场景
\uD83D\uDCD0 指令跟随与结构化输出能力强
支持明确输出格式（表格、步骤、要点列表）
适合 API 场景下对输出结构、字段、规范性要求高的业务系统
易与农业信息化平台、业务系统集成
\uD83D\uDEE0 开源可控，适合行业私有化部署
支持私有化部署与行业二次训练
可作为农业行业专用大模型或多模型体系中的核心底座
便于结合农业私有数据构建专属智能能力
典型应用场景
\uD83C\uDF31 农业生产与技术服务
作物种植方案咨询（播种、施肥、灌溉、采收）
病虫害识别与防治建议（基于文本/描述）
农业气象与灾害风险提示解读
\uD83D\uDCDA 农业知识问答与科普
农业技术问答系统
农技推广、农业培训智能助手
面向农户与基层人员的智能咨询服务
\uD83D\uDCC4 农业文本与内容生成
农业技术资料、方案说明编写
农业政策文件解读与摘要
农业报告、项目材料、申报文档辅助生成
\uD83C\uDFE2 农业管理与信息化系统
农业管理平台智能助手
农业数据解释与分析辅助
农业企业内部知识库（结合 RAG）
\uD83E\uDDE9 行业智能应用与平台集成
智慧农业平台
农业政务与监管系统
农业 SaaS / 农业大数据平台的 AI 能力模块
模型介绍
Qwen-Image-2512 是阿里巴巴通义千问团队推出的高性能文本生成图像（Text-to-Image）大模型。该模型在前代 Qwen-Image 的基础上进行了系统性升级，在人物真实感、自然场景细节、复杂语义理解以及文字渲染能力等方面表现突出，能够稳定地将自然语言描述转化为高质量、可商用的图像内容。
我们通过 API 形式对外提供 Qwen-Image-2512 服务，支持开发者与企业快速接入，实现自动化、高一致性的图像生成能力。
核心特点
1. 更强的真实感与细节表现
人物面部结构、肤质、光影更加自然
对自然场景、材质纹理、复杂构图的还原度显著提升
适合对视觉质量要求较高的应用场景
2. 优秀的文字生成与排版能力
能准确生成海报、封面、插画中的中英文文字
相比多数同类模型，文字可读性更高、变形更少
适合营销物料、宣传图、信息型图片生成
3. 强语义理解与指令跟随能力
能理解较长、复杂的 Prompt 描述
对风格、构图、主体关系的控制更稳定
适合精细化创作与批量生成场景
4. 通用性强，风格覆盖广
支持写实、插画、二次元、设计稿、概念艺术等多种风格
适配不同行业与创作需求
5. API 友好，易于工程化落地
标准化接口，便于快速集成
适合自动化流程、批量生成、在线服务场景
典型应用场景
\uD83C\uDFA8 内容创作与设计辅助
海报、封面、插画、配图生成
自媒体、内容平台的图像自动生产
设计师灵感草图与初稿生成
\uD83D\uDECD 电商与营销
商品展示图、场景图生成
营销海报、活动主视觉设计
多版本广告素材自动生成与测试
\uD83D\uDCF1 产品与应用集成
AI 绘画应用、创作类工具
文生图 SaaS 平台
企业内部内容生产系统
\uD83C\uDFAE 游戏与虚拟内容
游戏角色、场景概念图生成
虚拟人物、虚拟世界素材创作
剧情插图与世界观设定辅助
\uD83D\uDCCA 企业与行业应用
PPT / 报告配图自动生成
教育、培训、知识可视化
品牌视觉资产的规模化生产
MAI-UI-8B 模型详细介绍
MAI-UI-8B 是由 Tongyi-MAI 推出的 8B 参数级多模态 GUI 智能体基础模型，面向真实图形用户界面（UI）环境，具备从“理解界面”到“执行操作”的完整闭环能力，可直接驱动应用与系统完成复杂交互任务。
\uD83D\uDD11 主要模型特色
1. 面向真实 UI 的原生智能体能力
不同于仅做界面描述的多模态模型，MAI-UI-8B 原生支持 UI 操作语义建模，可将自然语言目标转化为可执行的界面动作序列（点击、输入、滑动、返回等）。
2. 视觉 + 语言深度融合的界面理解
模型可同时理解界面截图、控件布局与文本语义，实现精准的 UI 元素定位（UI Grounding），适用于动态界面与复杂交互流程。
3. 强化学习驱动的任务规划与执行
通过大规模真实/仿真环境中的在线强化学习训练，模型具备多步规划能力，能够在不确定界面状态下持续修正策略并完成目标任务。
4. 高性价比 8B 规模设计
在保持优秀 UI 操作准确率的同时，8B 参数规模兼顾推理效率与部署成本，适合端侧、边缘计算或私有化部署场景。
5. 开源可扩展，易于系统集成
模型已在 Hugging Face 开源，可作为 UI Agent、自动化系统或智能助手的核心推理引擎，方便与现有业务系统、工具链和 Agent 框架集成。
\uD83D\uDE80 典型应用场景
\uD83D\uDCF1 智能设备 / 手机自动化操作
根据自然语言指令自动完成 App 操作，如设置调整、信息填写、功能导航等，构建“可操作手机的 AI”。
\uD83E\uDDD1\uD83D\uDCBC 智能助手与个人 Agent
作为智能助理的执行层，帮助用户完成跨应用任务，例如查询信息、整理内容、发送消息等。
\uD83C\uDFE2 企业流程自动化（RPA + AI）
用于企业内部系统、后台管理界面或 Web 系统的自动操作，降低人工成本，提升流程效率。
\uD83E\uDDEA UI 测试与交互验证
自动执行界面操作路径，验证 UI 流程完整性与可用性，辅助测试与质量保障。
\uD83D\uDD27 端云协同智能系统
在端侧完成基础 UI 感知与操作，在云端进行复杂规划与策略优化，实现效率与能力平衡。
GLM-Image 是由智谱推出的高质量文本生成图像模型，面向真实生产场景设计，支持通过 API 将自然语言快速转化为具备良好语义一致性与视觉表现力的图像内容。
核心能力
高质量文生图：准确理解中文与英文提示词，生成结构清晰、细节丰富的图像
中文语义与文字渲染能力强：在包含中文文本、标识、海报元素等场景下表现稳定
复杂语义理解：支持多对象、多属性、多动作组合的描述生成
风格与场景泛化能力好：适配插画、写实、设计稿、概念图等多种视觉风格
API 友好：适合自动化调用，便于集成到现有业务系统和工作流中
API 使用优势
低接入成本：标准 REST API，快速完成模型集成
稳定可扩展：适合高并发、批量生成等生产级使用
参数灵活：支持通过提示词精细控制画面内容、风格与构图
适合平台化服务：可作为图像生成能力模块，对外提供 SaaS 或 PaaS 服务
典型应用场景
内容与营销：广告配图、社媒图片、活动海报自动生成
设计与创意：UI/视觉草图、概念设计、灵感探索
电商与零售：商品展示图、场景化陈列图生成
教育与出版：教材插图、科普配图、内容可视化
AIGC 平台能力底座：作为图像生成 API，服务下游多业务系统
\uD83E\uDDE0 模型简介：Qwen3-VL-Embedding-8B
Qwen3-VL-Embedding-8B 是基于强大的 Qwen3-VL 多模态基础模型构建的 8 B 参数向量嵌入模型，可将文本、图像、截图、视频及其任意组合输入编码为高质量语义向量，支持跨模态语义对齐与相似性比较。
⭐ 模型特点
\uD83D\uDD39 多模态支持
能够对 文本、图像、截图及混合模态（如文本+图像） 输入生成统一语义向量，打破单一模态检索限制。
\uD83D\uDD39 强语义表达
采用 Qwen3-VL 底层架构，生成的嵌入向量在语义层面具有更丰富、更细粒度的表达，提升跨模态相似性计算与匹配准确性。
\uD83D\uDD39 大规模参数（8B）
8 B 参数规模赋予模型更强的理解能力与语义建模能力，适合对复杂场景、多语言环境下的语义检索任务进行高效支持。
\uD83D\uDD39 多语言能力
支持 30 多种语言 的统一向量表示，可满足全球化检索与跨语言检索需求。
\uD83D\uDD39 可定制输出维度
向量维度最高可达 4096，并支持用户自定义输出维度（如 64–4096），方便根据任务需求优化存储与检索效率。
\uD83D\uDD39 超长上下文
可处理长文本或组合内容的上下文长度高达 32k tokens，适用于大段内容的相似性比较与长文检索场景。
\uD83D\uDE80 典型应用场景
\uD83D\uDD0D 多模态检索系统
在搜索引擎或知识库中构建跨模态检索 pipelines，将用户查询（文本或图像）转化为向量，与数据库中的向量进行相似度匹配，支持精准检索。
\uD83E\uDD16 搜索召回与排序增强
结合向量召回与重排序模型（如 Qwen3-VL-Reranker-8B）形成二阶段检索架构，提高检索结果排序质量与精确度。
\uD83D\uDCDA 聚类与组织内容
对大量文档、图像、视频等内容生成统一语义向量，可用于内容聚类、主题划分与信息组织。
\uD83E\uDDE0 知识增强生成（RAG）
作为 RAG（检索增强生成）系统的召回组件，提供高质量嵌入向量作为上下文输入，提高下游生成任务的回答准确性与相关性。
\uD83C\uDF10 跨语言语义检索
支持多语言语义表示，可以用于跨语种检索、翻译对齐、跨语义推荐等场景。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1150,"name":"32K","slug":"context_size_32k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-f8crosrdk-id-y4z8cy4z9-context_size.png","en_name":"32K","color":"blue","description":"上下文大小","href":"","created_at":1734693847615,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1196,"name":"向量化和重排","slug":"embedding-rerank","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-feature-extraction.svg","en_name":"Embedding and Rerank","color":"","description":"","href":"","created_at":1741323767491,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1223,"name":"4096 维","slug":"4096-dim","icon":"","en_name":"4096 dim","color":"","description":"","href":"","created_at":1749203697103,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":25754,"name":"Qwen3-VL-Embedding-8B","namespace":"hf-models","mirror_namespace":"Qwen","mirror_url":"https://huggingface.co/Qwen/Qwen3-VL-Embedding-8B"},"total_calls":21,"operation_summary":{"service_id":6392,"unit_tag":0,"min_price":0.0000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0000,"max_input_million_tokens_price":0.35000000,"max_output_million_tokens_price":0.35000000,"free_operation_count":2,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":602,"account":0,"account_type":0,"type":"embeddings","service_id":6392,"service_type":"serverless","model_id":25754,"vendor":"gitee","region":"prod-gitee-urc-ip","compute":"gitee/prod-gitee-urc-ip/serverless_api/gpu/large/default","name":"多模态向量(小批量)","description":"","path":"v1/embeddings","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"input","type":"string","control":"multimodal","category":"main","required":true,"max_length":25,"min_length":1,"description":"Embedding 输入内容，支持以数组形式传入多个文本或图片；图片支持 Base64 编码或 URL。"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1768890052000,"updated_at":1787905442000,"vendor_info":{"id":0,"name":"gitee","label":"Gitee","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8n73csd10-image 5.svg"},"region_info":{"name":"prod-gitee-urc-ip","label":"乌鲁木齐","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"gitee/prod-gitee-urc-ip/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"NVIDIA系列","gpu_memory_gb":24,"gpu_performance_rate":"full","region_name":"prod-gitee-urc-ip","vendor_name":"gitee","type":"SERVERLESS_API","storage":0},"ids":[602],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":809,"account":0,"account_type":0,"type":"embeddings","service_id":6392,"service_type":"serverless","model_id":25754,"vendor":"gitee","region":"prod-gitee-urc-ip","compute":"gitee/prod-gitee-urc-ip/serverless_api/gpu/large/default","name":"多模态向量(大批量)","description":"","path":"v1/embeddings","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0.35000000,"output_million_tokens_price":0.35000000,"input_million_tokens_price_average":0.35000000,"output_million_tokens_price_average":0.35000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"input","type":"string","control":"multimodal","category":"main","required":true,"max_length":1000,"min_length":26,"description":"Embedding 输入内容，支持以数组形式传入多个文本或图片；图片支持 Base64 编码或 URL。"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1783564510000,"updated_at":1787558577000,"vendor_info":{"id":0,"name":"gitee","label":"Gitee","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8n73csd10-image 5.svg"},"region_info":{"name":"prod-gitee-urc-ip","label":"乌鲁木齐","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"gitee/prod-gitee-urc-ip/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"NVIDIA系列","gpu_memory_gb":24,"gpu_performance_rate":"full","region_name":"prod-gitee-urc-ip","vendor_name":"gitee","type":"SERVERLESS_API","storage":0},"ids":[809],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":1216}]},{"service":{"id":6395,"ident":"Qwen3-VL-Reranker-8B","type":"serverless","vendor":"","region":"","compute":"","model":25752,"name":"Qwen3-VL-Reranker-8B","remark":"Qwen3-VL-Reranker-8B 是一款基于 Qwen3-VL 的 8B 参数多模态重排序模型，用于对文本、图像等检索结果进行高精度相关性排序。","price":0.0000,"props":{"name":"Qwen3-VL-Reranker-8B","service_type":"serverless"},"inventories":0,"created_at":1768895312000,"updated_at":1768896022000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3-VL-Reranker-8B 是基于 Qwen3-VL 多模态大模型打造的高性能重排序（Reranker）模型，支持对文本、图像、截图等多模态候选结果进行精准相关性评分。该模型通常与向量检索（Embedding）模型配合使用，在召回阶段之后对候选结果进行二次精排，显著提升检索结果的准确性与用户体验。
模型特点
多模态重排序能力：统一建模文本与视觉信息，准确理解跨模态语义相关性
高精度精排：针对 Top-K 候选结果进行细粒度打分，显著优于仅依赖向量相似度
大参数规模（8B）：具备更强的语义理解与复杂场景判别能力
检索系统友好：适合与现有搜索、RAG、向量数据库无缝集成
应用场景
多模态搜索引擎：文本搜图、图文混合搜索、视频内容检索
RAG（检索增强生成）：提升知识检索阶段的上下文相关性与答案质量
智能推荐与排序：对候选内容进行精准相关性排序
企业级内容检索：文档、图片、设计稿、截图等内部知识库精排
\uD83E\uDDE0 模型概述
FLUX.2-klein-9B 是 Black Forest Labs 旗下 FLUX.2 [klein] 系列的旗舰小型图像生成与编辑模型。整个模型设计侧重于高质量输出与非常低的推理延迟之间的最佳平衡，尤其适合实时交互式场景。
该模型使用 FLUX Non-Commercial License v2.1 许可证。
\uD83D\uDCCC 使用场景
\uD83D\uDD8C 创意与艺术创作
实时生成高质量图像用于插画、海报、概念艺术等。
根据文本提示快速试验视觉主题和构图。
适合创意设计师、艺术家、视觉内容创作者。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1148,"name":"图像生成与处理","slug":"image_generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-to-image.svg","en_name":"Image Generation","color":"","description":"","href":"","created_at":1734660944141,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":25731,"name":"FLUX.2-klein-9B","namespace":"hf-models","mirror_namespace":"black-forest-labs","mirror_url":"https://huggingface.co/black-forest-labs/FLUX.2-klein-9B"},"total_calls":27,"operation_summary":{"service_id":6468,"unit_tag":0,"min_price":0.0050,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.1000,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":3},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":651,"account":0,"account_type":0,"type":"text2image","service_id":6468,"service_type":"serverless","model_id":25731,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"文生图(1024*1024)","description":"","path":"v1/images/generations","protocol":"http","price":0.0050,"average_price":0.0050,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":1,"params":[{"in":"body","name":"size","type":"string","control":"select","options":[{"label":"1:1 (256*256)","value":"256x256"},{"label":"1:1 (512*512)","value":"512x512"},{"label":"1:1 (1024*1024)","value":"1024x1024"},{"label":"4:3 (1024*768)","value":"1024x768"},{"label":"3:4 (768*1024)","value":"768x1024"},{"label":"16:9 (1024*576)","value":"1024x576"},{"label":"9:16 (576*1024)","value":"576x1024"},{"label":"3:2 (1024*640)","value":"1024x640"},{"label":"2:3(640*1024)","value":"640x1024"}],"required":false,"description":"图片尺寸","default_value":"1024x1024"},{"in":"body","name":"prompt","type":"string","control":"string","samples":[{"label":"机器人","value":"A robot sitting on open grassland, painting on a canvas."},{"label":"女孩","value":"A cute Japanese girl, full frontal body shot, sitting in the front passenger seat of a car, wearing a seatbelt, shot with a Sony α7R IV and cinematic lighting, surrealist photography."},{"label":"卡通角色","value":"An illustration of a cartoon samurai wearing a kimono, front view, full body, with floral patterns, 2D symmetrical composition."}],"category":"main","required":true,"description":"生成图片的提示词，只支持英文。"},{"in":"body","name":"width","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的宽度，最大为1024，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"height","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的高度， 最大为1024，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"num_inference_steps","step":1,"type":"number","control":"slider","maximum":50,"minimum":1,"required":false,"description":"表示图像生成时的迭代步数，数值越大图像越清晰，但生成时间也越长。","default_value":4},{"in":"body","name":"guidance_scale","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"提示词引导系数，增大数值将引导生成图片的过程更加遵循提示词。","default_value":1},{"in":"body","name":"seed","type":"integer","control":"number","required":false,"description":"用于确保随机过程的结果可复现，从而保证实验的一致性。"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1773900888000,"updated_at":1782712895000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[651],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":653,"account":0,"account_type":0,"type":"image2image","service_id":6468,"service_type":"serverless","model_id":25731,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"单图生图(1024*1024)","description":"","path":"v1/images/edits","protocol":"http","price":0.0500,"average_price":0.0500,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":6,"params":[{"in":"form","name":"size","type":"string","control":"select","options":[{"label":"1:1 (256*256)","value":"256x256"},{"label":"1:1 (512*512)","value":"512x512"},{"label":"1:1 (1024*1024)","value":"1024x1024"},{"label":"4:3 (1024*768)","value":"1024x768"},{"label":"3:4 (768*1024)","value":"768x1024"},{"label":"16:9 (1024*576)","value":"1024x576"},{"label":"9:16 (576*1024)","value":"576x1024"},{"label":"3:2 (1024*640)","value":"1024x640"},{"label":"2:3(640*1024)","value":"640x1024"}],"required":false,"description":"图片尺寸","default_value":"1024x1024"},{"in":"form","name":"width","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的宽度，最大为1024，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"form","name":"height","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的高度， 最大为1024，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"form","name":"image","type":"array","items":{"type":"file","control":"file"},"props":{"accept":".jpg, .jpeg, .png, .gif, .webp","textField":false,"imageField":{"max_length":1,"show_array":true}},"accept":".jpg, .jpeg, .png, .gif, .webp","control":"multimodal","example":["/path/to/image.png"],"samples":[{"label":"示例","value":[{"type":"images","value":["https://gitee.com/gitee-ai/moark-assets/raw/master/pytest_image/woman1.webp"]}]}],"category":"main","required":false,"max_length":1,"min_length":1,"description":"图片"},{"in":"form","name":"mask","type":"file","label":"参考图","control":"file","visible":false,"category":"main","required":false,"description":"用于引导生成预期图像的描述文字。"},{"in":"form","name":"prompt","type":"string","control":"string","samples":[{"label":"示例","value":"A woman is gently petting a cat, with a bird standing beside them in the same scene."}],"category":"main","required":true,"description":"需要生成图片的文案说明"},{"in":"form","name":"num_inference_steps","step":1,"type":"integer","control":"slider","maximum":50,"minimum":1,"required":false,"description":"生成图像时模型执行迭代的次数。数值越大，生成的图像质量会更高","default_value":4},{"in":"form","name":"guidance_scale","step":1,"type":"number","control":"slider","maximum":10,"minimum":1,"required":false,"description":"调节模型在生成图像时对输入提示的遵从程度，数值越大，遵从越严格","default_value":1},{"in":"body","name":"seed","type":"integer","control":"number","required":false,"description":"用于确保随机过程的结果可复现，从而保证实验的一致性。"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1773900977000,"updated_at":1782712882000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[653],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":655,"account":0,"account_type":0,"type":"image2image","service_id":6468,"service_type":"serverless","model_id":25731,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"多图生图(1024*1024)","description":"","path":"v1/images/edits","protocol":"http","price":0.1000,"average_price":0.1000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":10,"params":[{"in":"form","name":"size","type":"string","control":"select","options":[{"label":"1:1 (256*256)","value":"256x256"},{"label":"1:1 (512*512)","value":"512x512"},{"label":"1:1 (1024*1024)","value":"1024x1024"},{"label":"4:3 (1024*768)","value":"1024x768"},{"label":"3:4 (768*1024)","value":"768x1024"},{"label":"16:9 (1024*576)","value":"1024x576"},{"label":"9:16 (576*1024)","value":"576x1024"},{"label":"3:2 (1024*640)","value":"1024x640"},{"label":"2:3(640*1024)","value":"640x1024"}],"required":false,"description":"图片尺寸","default_value":"1024x1024"},{"in":"form","name":"width","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的宽度，最大为1024，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"form","name":"height","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的高度， 最大为1024，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"form","name":"image","type":"array","items":{"type":"file","control":"file"},"props":{"accept":".jpg, .jpeg, .png, .gif, .webp","textField":false,"imageField":{"max_length":3}},"accept":".jpg, .jpeg, .png, .gif, .webp","control":"multimodal","example":["/path/to/image.png"],"samples":[{"label":"示例","value":[{"type":"images","value":["https://gitee-ai.su.bcebos.com/samples/images/flux1.png","https://gitee-ai.su.bcebos.com/samples/images/flux2.png","https://gitee-ai.su.bcebos.com/samples/images/flux3.png"]}]}],"category":"main","required":false,"max_length":3,"min_length":2,"description":"图片"},{"in":"form","name":"mask","type":"file","label":"参考图","control":"file","visible":false,"category":"main","required":false,"description":"用于引导生成预期图像的描述文字。"},{"in":"form","name":"prompt","type":"string","control":"string","samples":[{"label":"示例","value":"The person from image 1 is petting the cat from image 2, the bird from image 3 is next to them."}],"category":"main","required":true,"description":"需要生成图片的文案说明"},{"in":"form","name":"num_inference_steps","step":1,"type":"integer","control":"slider","maximum":50,"minimum":1,"required":false,"description":"生成图像时模型执行迭代的次数。数值越大，生成的图像质量会更高","default_value":4},{"in":"form","name":"guidance_scale","step":1,"type":"number","control":"slider","maximum":10,"minimum":1,"required":false,"description":"调节模型在生成图像时对输入提示的遵从程度，数值越大，遵从越严格","default_value":1},{"in":"body","name":"seed","type":"integer","control":"number","required":false,"description":"用于确保随机过程的结果可复现，从而保证实验的一致性。"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1773901073000,"updated_at":1782712913000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[655],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":6475,"ident":"search","type":"serverless","vendor":"","region":"","compute":"","model":0,"name":"全网搜索","remark":"从近百亿网页和生态内容源中搜索高质量世界知识，例如新闻、图片、百科、文库等。","price":0.0000,"props":{"service_type":"serverless"},"inventories":0,"created_at":1764062786000,"updated_at":1779962254000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-2jn01ns1t-20260528-161814.png","description_html":"面向 AI 应用与智能检索场景的世界知识搜索接口，覆盖近百亿级网页及多生态内容源，包括新闻、图片、百科、文库、论坛与公开资料等。支持语义级搜索、高质量结果召回与结构化知识提取，可作为 Agent、RAG、Copilot 与内容分析系统的外部知识入口。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1277,"name":"模型生态扩展","slug":"model_ecosystem_expansion","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-ag4axdnzk-plug.svg","en_name":"Model Ecosystem Expansion","color":"","description":"为大模型赋予实时联网、数据查询等拓展能力，轻松构建强大的 Agent 应用。","href":"","created_at":1779957002588,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1278,"name":"联网搜索","slug":"online_search","icon":"","en_name":"Online Search","color":"","description":"","href":"","created_at":1779957560580,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"total_calls":7,"operation_summary":{"service_id":6475,"unit_tag":0,"min_price":0.0200,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0300,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":4},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":722,"account":0,"account_type":0,"type":"web_search_v2","service_id":6475,"service_type":"serverless","model_id":0,"vendor":"baidu-cloud","region":"prod-bce-default-ip","compute":"baidu-cloud/prod-bce-default-ip/serverless_api/gpu/large/default","name":"web-search v2","description":"百度搜索原生格式接口","path":"v1/web-search-v2","protocol":"http","price":0.0200,"average_price":0.0200,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":1,"params":[{"in":"body","name":"content","type":"string","label":"搜索内容","control":"text","category":"main","required":true,"description":"搜索查询内容"},{"in":"body","name":"search_recency_filter","type":"string","label":"搜索时间范围","control":"select","options":[{"label":"一周内","value":"week"},{"label":"一个月内","value":"month"},{"label":"半年内","value":"semiyear"},{"label":"一年内","value":"year"}],"category":"extra","required":false,"description":"根据网页发布时间进行筛选"},{"in":"body","name":"search_site_filter","type":"string","label":"搜索站点范围","control":"string-array","category":"extra","required":false,"max_length":100,"description":"指定站点搜索条件，仅在设置的站点中进行内容搜索，每行一个站点","placeholder":"tieba.baidu.com\nbaike.baidu.com"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1778469538000,"updated_at":1779965444000,"vendor_info":{"id":0,"name":"baidu-cloud","label":"百度智能云","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-w36p9kcyq-3412a51e.png"},"region_info":{"name":"prod-bce-default-ip","label":"百度智能云默认","props":{}},"compute_info":{"id":"baidu-cloud/prod-bce-default-ip/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"prod-bce-default-ip","vendor_name":"baidu-cloud","type":"SERVERLESS_API","storage":0},"ids":[722],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":0},{"id":723,"account":0,"account_type":0,"type":"web_search","service_id":6475,"service_type":"serverless","model_id":0,"vendor":"baidu-cloud","region":"prod-bce-default-ip","compute":"baidu-cloud/prod-bce-default-ip/serverless_api/gpu/large/default","name":"web-search","description":"","path":"v1/web-search","protocol":"http","price":0.0200,"average_price":0.0200,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":1,"params":[{"in":"body","name":"count","type":"number","control":"number","visible":false,"category":"main","required":false,"description":"返回的搜索结果数量（1-8），默认为8","default_value":8}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1763023920000,"updated_at":1779965453000,"vendor_info":{"id":0,"name":"baidu-cloud","label":"百度智能云","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-w36p9kcyq-3412a51e.png"},"region_info":{"name":"prod-bce-default-ip","label":"百度智能云默认","props":{}},"compute_info":{"id":"baidu-cloud/prod-bce-default-ip/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"prod-bce-default-ip","vendor_name":"baidu-cloud","type":"SERVERLESS_API","storage":0},"ids":[723],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":0},{"id":908,"account":0,"account_type":0,"type":"web_search_custom","service_id":6475,"service_type":"serverless","model_id":0,"vendor":"baidu-cloud","region":"prod-bce-default-ip","compute":"baidu-cloud/prod-bce-default-ip/serverless_api/gpu/large/default","name":"volc web-search custom","description":"豆包搜索 Custom 版，路径 /v1/web-search/custom，使用平台 Web 搜索请求响应格式","path":"v1/web-search/custom","protocol":"http","price":0.0300,"average_price":0.0300,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":1,"params":[{"in":"body","name":"query","type":"string","label":"搜索内容","control":"text","category":"main","required":true,"description":"搜索关键字或语句，最多 100 个字符"},{"in":"body","name":"count","type":"number","label":"结果数量","control":"number","maximum":50,"minimum":1,"category":"extra","required":false,"description":"返回的网页结果数量，1-50，默认 10","default_value":10},{"in":"body","name":"search_recency_filter","type":"string","label":"发布时间","control":"select","options":[{"label":"一天内","value":"day"},{"label":"一周内","value":"week"},{"label":"一个月内","value":"month"},{"label":"一年内","value":"year"}],"category":"extra","required":false,"description":"根据网页发布时间筛选"},{"in":"body","name":"search_site_filter","type":"string","label":"指定站点","control":"string-array","category":"extra","required":false,"max_length":20,"description":"仅搜索指定站点，每行一个域名，最多 20 个","placeholder":"example.com\nnews.example.com"},{"in":"body","name":"block_hosts","type":"string","label":"屏蔽站点","control":"string-array","category":"extra","required":false,"max_length":5,"description":"屏蔽指定站点，每行一个域名，最多 5 个","placeholder":"example.com"},{"in":"body","name":"auth_info_level","type":"number","label":"权威度","control":"select","options":[{"label":"不限制","value":"0"},{"label":"仅非常权威","value":"1"}],"category":"extra","required":false,"description":"限制搜索结果的权威等级","default_value":0},{"in":"body","name":"industry","type":"string","label":"行业搜索","control":"select","options":[{"label":"金融","value":"finance"},{"label":"电子游戏","value":"game"},{"label":"政府及权威媒体","value":"gov"}],"category":"extra","required":false,"description":"按行业范围执行搜索"},{"in":"body","name":"content_formats","type":"string","label":"正文格式","control":"select","options":[{"label":"纯文本","value":"text"},{"label":"Markdown","value":"markdown"}],"category":"extra","required":false,"description":"指定返回正文的格式","default_value":"text"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1786428263000,"updated_at":1786535177000,"vendor_info":{"id":0,"name":"baidu-cloud","label":"百度智能云","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-w36p9kcyq-3412a51e.png"},"region_info":{"name":"prod-bce-default-ip","label":"百度智能云默认","props":{}},"compute_info":{"id":"baidu-cloud/prod-bce-default-ip/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"prod-bce-default-ip","vendor_name":"baidu-cloud","type":"SERVERLESS_API","storage":0},"ids":[908],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Volcengine"],"visible":true,"unit_tag":0},{"id":909,"account":0,"account_type":0,"type":"web_search_global","service_id":6475,"service_type":"serverless","model_id":0,"vendor":"baidu-cloud","region":"prod-bce-default-ip","compute":"baidu-cloud/prod-bce-default-ip/serverless_api/gpu/large/default","name":"volc web-search global","description":"豆包搜索 Global 版，路径 /v1/web-search/global，使用平台 Web 搜索请求响应格式","path":"v1/web-search/global","protocol":"http","price":0.0300,"average_price":0.0300,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":1,"params":[{"in":"body","name":"query","type":"string","label":"搜索内容","control":"text","category":"main","required":true,"description":"搜索关键字或语句，最多 100 个字符"},{"in":"body","name":"count","type":"number","label":"结果数量","control":"number","maximum":20,"minimum":1,"category":"extra","required":false,"description":"返回的搜索结果数量，1-20，默认 10","default_value":10},{"in":"body","name":"max_snippet_length","type":"number","label":"摘要长度","control":"number","maximum":3000,"minimum":1,"category":"extra","required":false,"description":"单个摘要片段的最大 tokens，1-3000","default_value":500},{"in":"body","name":"max_image_count_per_doc","type":"number","label":"单条结果图片数","control":"number","maximum":10,"minimum":1,"category":"extra","required":false,"description":"单条搜索结果最多返回的图片数量，1-10","default_value":3}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1786428263000,"updated_at":1786535269000,"vendor_info":{"id":0,"name":"baidu-cloud","label":"百度智能云","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-w36p9kcyq-3412a51e.png"},"region_info":{"name":"prod-bce-default-ip","label":"百度智能云默认","props":{}},"compute_info":{"id":"baidu-cloud/prod-bce-default-ip/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"prod-bce-default-ip","vendor_name":"baidu-cloud","type":"SERVERLESS_API","storage":0},"ids":[909],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Volcengine"],"visible":true,"unit_tag":0}]},{"service":{"id":6519,"ident":"Qwen3-VL-Embedding-2B","type":"serverless","vendor":"","region":"","compute":"","model":25753,"name":"Qwen3-VL-Embedding-2B","remark":"Qwen3-VL-Embedding-2B 是一个支持文本、图像等多模态输入、用于生成统一高维语义向量的 2 B 参数级多模态嵌入模型，常用于跨模态检索和相似性搜索等任务。","price":0.0000,"props":{"name":"Qwen3-VL-Embedding-2B","service_type":"serverless"},"inventories":0,"created_at":1769411304000,"updated_at":1769411558000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"\uD83D\uDCCC 模型简介
Qwen3-VL-Embedding-2B 是阿里 Qwen 系列最新的多模态嵌入模型，基于强大的 Qwen3-VL 多模态基础架构构建。它通过一个统一的向量空间将文本、图像等不同类型的数据映射成高维语义向量，使得跨模态理解与相似性计算变得高效可靠。
\uD83D\uDD0D 主要特点
⭐ 多模态支持
同时支持文本、图片、截图等多种输入格式，并可组合混合模态输入（例如“图文混合”）。
\uD83E\uDDE0 统一语义空间
生成的向量在一个共享空间内有效表达语义信息，使得跨模态相似性计算和搜索成为可能。
\uD83D\uDCCF 灵活的向量维度
默认输出维度可达 2048 维，并支持根据不同场景自定义输出维度（64–2048）。
\uD83C\uDF0D 多语言能力
支持 30 多种语言，适用于全球化应用。
\uD83E\uDDE9 可定制指令
支持为不同任务添加“指令”，提升下游任务表现。
\uD83D\uDE80 高效部署
支持量化处理（Quantization）来减少模型体积与推理延迟，便于在 API/服务端大规模部署。
\uD83D\uDCDA 大上下文理解能力
最大可处理 32K token 的输入，上下文覆盖能力强，适合长文档场景。
\uD83E\uDDE0 主要使用场景
\uD83D\uDD0D 1) 跨模态搜索与检索 API
构建支持“图文 + 文本”检索的搜索 API，例如：
图片内容检索
多模态语义搜索
文本提示下的视觉相似性搜索
这些场景中，通过计算向量相似度即可迅速检索相关结果。
\uD83D\uDCC2 2) 相似度计算与聚类
适用于分析不同模态数据的语义相似性，用于：
用户行为聚类
内容推荐系统
语义聚类与分类服务
\uD83D\uDCCC 3) 数据索引与嵌入服务
为文档、图像生成统一嵌入，用于后端向量数据库（如 Milvus、Pinecone）索引，架构向量搜索 API。
\uD83E\uDDE0 4) 多模态智能问答与理解服务
结合向量搜索和大模型可用于：
图像理解问答（Visual QA）
复杂场景理解与推理扩展（通过向量检索 + 下游模型组合）
\uD83D\uDD04 5) 细粒度重排序（Re-ranking）（配合 Reranker）
在搜索场景中可配合 Qwen3-VL-Reranker 进行检索结果精排，形成“粗排 + 精排”检索 API 管线，提升准确性与用户体验。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1163,"name":"多模态","slug":"multi-modal","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287910629,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1196,"name":"向量化和重排","slug":"embedding-rerank","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-feature-extraction.svg","en_name":"Embedding and Rerank","color":"","description":"","href":"","created_at":1741323767491,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1225,"name":"2048 维","slug":"2048-dim","icon":"","en_name":"2048 dim","color":"","description":"","href":"","created_at":1751282794427,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":25753,"name":"Qwen3-VL-Embedding-2B","namespace":"hf-models","mirror_namespace":"Qwen","mirror_url":"https://huggingface.co/Qwen/Qwen3-VL-Embedding-2B"},"total_calls":10,"operation_summary":{"service_id":6519,"unit_tag":0,"min_price":0.0000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0000,"max_input_million_tokens_price":0.09000000,"max_output_million_tokens_price":0.09000000,"free_operation_count":2,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":610,"account":0,"account_type":0,"type":"embeddings","service_id":6519,"service_type":"serverless","model_id":25753,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"多模态向量(小批量)","description":"","path":"v1/embeddings","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"input","type":"string","control":"multimodal","category":"main","required":true,"max_length":25,"min_length":1,"description":"输入内容"},{"in":"body","name":"dimensions","type":"integer","control":"number","maximum":2048,"minimum":64,"required":false,"description":"维度：返回的嵌入向量的维度（dimension），即每个文本将被转换为多少维的向量。","default_value":1024},{"in":"body","name":"instruction","type":"string","control":"string","required":false,"description":"用于指示模型按照什么标准对候选项进行向量化"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1769411304000,"updated_at":1787905796000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[610],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":810,"account":0,"account_type":0,"type":"embeddings","service_id":6519,"service_type":"serverless","model_id":25753,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"多模态向量(大批量)","description":"","path":"v1/embeddings","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0.09000000,"output_million_tokens_price":0.09000000,"input_million_tokens_price_average":0.09000000,"output_million_tokens_price_average":0.09000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"input","type":"string","control":"multimodal","category":"main","required":true,"max_length":1000,"min_length":26,"description":"输入内容"},{"in":"body","name":"dimensions","type":"integer","control":"number","maximum":2048,"minimum":64,"required":false,"description":"维度：返回的嵌入向量的维度（dimension），即每个文本将被转换为多少维的向量。","default_value":1024},{"in":"body","name":"instruction","type":"string","control":"string","required":false,"description":"用于指示模型按照什么标准对候选项进行向量化"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1783564510000,"updated_at":1787558566000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[810],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":1216}]},{"service":{"id":6521,"ident":"LTX-2","type":"serverless","vendor":"","region":"","compute":"","model":25700,"name":"LTX-2","remark":"LTX-2 是首个基于 DiT 的音视频基础模型，集成了现代视频生成的所有核心能力：同步音视频、高保真度、多种性能模式、生产就绪输出、API 访问和开放访问。","price":0.0000,"props":{"name":"LTX-2","service_type":"serverless"},"inventories":0,"created_at":1769413396000,"updated_at":1769416433000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai-testing.su.bcebos.com/v1/uploads/img/id-1m153md4z-1652783139615-628375426db5127097cf5442.webp","description_html":"LTX-2 是首个基于 DiT 的音视频基础模型，集成了现代视频生成的所有核心能力：同步音视频、高保真度、多种性能模式、生产就绪输出、API 访问和开放访问。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":118,"name":"其他","slug":"other","icon":"","en_name":"Other","color":"","description":"","href":"","created_at":1700443952785,"sort_order":0,"categories":[{"id":4,"parent_id":0,"name":"许可证","slug":"license","created_at":1700107824458,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/license.svg","sort_order":4,"en_name":""}]},{"id":1191,"name":"视频生成","slug":"video-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text_to_video.svg","en_name":"Video Generation","color":"","description":"","href":"","created_at":1739973857876,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":25700,"name":"LTX-2","namespace":"hf-models","mirror_namespace":"Lightricks","mirror_url":"https://huggingface.co/Lightricks/LTX-2"},"total_calls":28,"operation_summary":{"service_id":6521,"unit_tag":1264,"min_price":0.3000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.3000,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":660,"account":0,"account_type":0,"type":"image2video","service_id":6521,"service_type":"serverless","model_id":25700,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"图生视频","description":"","path":"v1/async/videos/image-to-video","protocol":"http","price":0.3000,"average_price":0.3000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"form","name":"num_inference_steps","step":1,"type":"integer","control":"slider","maximum":40,"minimum":1,"required":false,"description":"生成视频时模型执行迭代的次数。数值越大，生成的质量会更高","default_value":8},{"in":"form","name":"num_frames","step":1,"type":"integer","control":"slider","maximum":241,"minimum":1,"required":false,"description":"生成视频的最大帧数。数值越大，生成的时间越长","default_value":73},{"in":"form","name":"fps","step":1,"type":"integer","control":"slider","maximum":24,"minimum":16,"required":false,"description":"用来控制视频的流畅度，数值越大画面越顺，但会更耗算力","default_value":24},{"in":"form","name":"enhance_prompt","type":"string","control":"text","required":false,"description":"是否对原始提示词进行自动优化与增强，以提升模型理解能力和最终生成效果。"},{"in":"form","name":"guidance_scale","step":1,"type":"number","control":"slider","maximum":10,"minimum":0,"required":false,"description":"调节模型在生成时对输入提示的遵从程度，数值越大，遵从越严格","default_value":1},{"in":"form","name":"seed","type":"integer","control":"number","required":false,"description":"用于确保随机过程的结果可复现，从而保证实验的一致性。"},{"in":"form","name":"lora_weights","type":"array","control":"array","required":false,"max_length":4,"min_length":0,"description":"lora权重配置"},{"in":"form","name":"lora_weights.url","type":"string","control":"text","required":false,"description":"lora模型权重的url，若为空则视为基础sdxl模型的推理。"},{"in":"form","name":"lora_weights.weight","type":"number","control":"number","required":false,"description":"若存在多个lora，此处填写长度须与lora数量相同，否则报错。这些数值用于不同lora权重之间的加权融合。","default_value":1},{"in":"form","name":"lora_scale","type":"number","control":"number","required":false,"description":"融合的lora权重对基础模型的影响程度。"},{"in":"form","name":"height","step":1,"type":"integer","control":"slider","maximum":2048,"minimum":8,"required":false,"description":"高度","default_value":512},{"in":"form","name":"width","step":1,"type":"integer","control":"slider","maximum":2048,"minimum":8,"required":false,"description":"宽度","default_value":512}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1774608073000,"updated_at":1782713590000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[660],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":1264}]},{"service":{"id":6526,"ident":"Qwen3-VL-Reranker-2B","type":"serverless","vendor":"","region":"","compute":"","model":25751,"name":"Qwen3-VL-Reranker-2B","remark":"Qwen3-VL-Reranker-2B 是一个基于 Qwen3-VL 多模态基础模型的 2 B 参数级别、多模态重新排序（Reranker）模型，用于对检索到的文本/图像/视频等候选结果进行精细相关性评分，从而提升多模态检索管道的准确性。","price":0.0000,"props":{"name":"Qwen3-VL-Reranker-2B","service_type":"serverless"},"inventories":0,"created_at":1769425743000,"updated_at":1769425779000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"\uD83E\uDDE0 模型特点
多模态统一建模：支持对文本、图像等多模态候选结果进行统一语义理解与相关性评估，实现跨模态精细化排序。
高精度相关性评分：具备良好的语义对齐与细粒度匹配能力，可准确区分候选结果之间的相关度差异。
高效轻量设计：2B 参数规模，在保证排序效果的同时兼顾推理效率，适合在线服务与高并发场景部署。
易集成精排组件：可通过 API 方式快速接入现有检索、推荐与 RAG 流程，作为精排模块提升整体系统效果。
\uD83D\uDCCC 使用场景
多模态搜索与推荐：图文混合搜索、以图搜文、以文搜图等场景中的结果重排与优选。
多模态 RAG 系统：对初步检索得到的多模态候选内容进行二次筛选与排序，提高最终生成结果的相关性。
商品与素材检索：在电商、图库、视频与设计素材平台中，对多模态候选内容进行相关性评分与排序。
知识库问答与智能客服：从多模态候选答案中选取最优匹配结果，用于提升问答准确率与用户体验。
\uD83D\uDCCC 模型概述
Qwen-Image-Layered 是由 Qwen 团队发布的开源图像分层分解模型（Apache-2.0 许可），其核心功能是 将一张普通的 RGB 图像自动分解成多个语义独立的 RGBA 图层，从而实现对图像内容的高保真、结构化和可控编辑。
⭐ 模型特色
\uD83E\uDDF1 1. 语义化图像分层（Semantic RGBA Decomposition）
模型通过先进的 AI 推理把语义和结构分离成多个独立的 RGBA（包含透明通道）层，每层通常表示一个明显的视觉组件（如主体、背景、文字、物件等）。
\uD83D\uDEE0 2. 原生可编辑性（Inherent Editability）
将图像分层后，每个层都可以独立编辑，避免传统图像编辑中常见的像素交叠、边缘污染等问题。典型支持的操作包括：
✅ 重新着色和调色
✅ 移动图像元素
✅ 缩放和旋转
✅ 删除不需要的对象
✅ 替换元素（例如把层内人物换成其他图像内容）
这些操作只影响目标图层，不会干扰其它图层的内容。
\uD83D\uDD01 3. 可变层数与递归分解
模型支持 可调层数，可根据图像内容复杂度生成少量或更多层；此外，任何一个已生成的层都可以 递归再分解，用于更精细的后续处理。
\uD83C\uDFAF 典型应用场景
\uD83C\uDFA8 1. 专业图像编辑与设计
为插画、宣传图、海报等生成可独立编辑的图层
快速调整视觉元素组合，无需手工分离像素
适用于设计师在日常创意工作中提高效率。
\uD83D\uDECD 2. 电商视觉资产生成
自动提取商品与背景
生成不同背景、颜色或组合用于电商展示与 A/B 广告测试
帮助电商团队快速扩展视觉资产池。
\uD83D\uDCE2 3. 内容创作与营销素材优化
批量为社交媒体、广告图、视觉素材拆层再编辑
生成多版本输出用于不同平台规格
提升内容创建的灵活性与生产效率。
\uD83C\uDFD7 4. 设计工作流程集成
与现有设计工具如 Photoshop/PPT 等集成
支持导出 PSD 或分层 PNG 等格式用于后续细粒度处理
适合设计团队和创意制作流水线自动化。
\uD83D\uDDBC 5. 图像增强与重构任务
对遮挡区域进行智能背景补全
分离出图像细节便于生成多视图或特定元素聚焦图
在图像分析与 AI 自动修复任务中也有潜力应用。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","stop_at":1789055940000,"sale_mode":"one_time","tags":[{"id":1148,"name":"图像生成与处理","slug":"image_generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-to-image.svg","en_name":"Image Generation","color":"","description":"","href":"","created_at":1734660944141,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1271,"name":"图像分层","slug":"image_layer","icon":"","en_name":"Image Layer","color":"","description":"","href":"","created_at":1769500774922,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":24326,"name":"Qwen-Image-Layered","namespace":"hf-models","mirror_namespace":"Qwen","mirror_url":"https://huggingface.co/Qwen/Qwen-Image-Layered"},"total_calls":24,"operation_summary":{"service_id":6551,"unit_tag":1270,"min_price":0.0500,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0500,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":615,"account":0,"account_type":0,"type":"image2image","service_id":6551,"service_type":"serverless","model_id":24326,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"图像分层","description":"图像分层","path":"v1/images/layers","protocol":"http","price":0.0500,"average_price":0.0500,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"form","name":"layers","step":1,"type":"number","control":"slider","maximum":4,"minimum":1,"required":true,"description":"输出图层数量","default_value":2},{"in":"form","name":"num_inference_steps","step":1,"type":"integer","control":"slider","maximum":50,"minimum":1,"required":false,"description":"生成图像时模型执行迭代的次数。数值越大，生成的图像质量会更高","default_value":15},{"in":"form","name":"true_cfg_scale","step":1,"type":"integer","control":"slider","maximum":10,"minimum":1,"required":false,"description":"控制模型内部约束强度，影响分层结果的稳定性与一致性。","default_value":4},{"in":"form","name":"negative_prompt","type":"string","control":"string","required":false,"description":"用于指定需要避免的内容特征，引导模型在生成或处理结果中尽量排除这些因素。"},{"in":"body","name":"seed","type":"integer","control":"number","required":false,"description":"用于确保随机过程的结果可复现，从而保证实验的一致性。","default_value":0}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1769500617000,"updated_at":1782705075000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[615],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":1270}]},{"service":{"id":6556,"ident":"Kimi-K2.5","type":"serverless","vendor":"","region":"","compute":"","model":25827,"name":"Kimi-K2.5","remark":"Kimi K2.5 是一个开源的原生多模态代理模型，通过在 Kimi-K2-Base 上约 15 万亿个混合视觉和文本令牌上持续预训练构建。它无缝将视觉和语言理解与先进的代理能力、即时和思考模式，以及会话和代理范式相结合。","price":0.0000,"props":{"name":"Kimi-K2.5","service_type":"serverless"},"inventories":0,"created_at":1769505011000,"updated_at":1787316799000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-2pzepojrt-id-g6laj4iw1-下载.png","description_html":"Kimi K2.5 是一个开源的原生多模态代理模型，通过在 Kimi-K2-Base 上约 15 万亿个混合视觉和文本令牌上持续预训练构建。它无缝将视觉和语言理解与先进的代理能力、即时和思考模式，以及会话和代理范式相结合。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":30,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1147,"name":"视觉模型","slug":"vision-language","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-vision-language.svg","en_name":"Vision Language","color":"","description":"","href":"","created_at":1734083179939,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1169,"name":"视觉","slug":"vision-language-label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735288027632,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1228,"name":"256K","slug":"context_size_256k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-9zvdx4rar-id-zd1snxypr-context_size.png","en_name":"256K","color":"blue","description":"上下文大小","href":"","created_at":1753329137839,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":25827,"name":"Kimi-K2.5","namespace":"hf-models","mirror_namespace":"moonshotai","mirror_url":"https://huggingface.co/moonshotai/Kimi-K2.5"},"total_calls":303,"operation_summary":{"service_id":6556,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":4.00000000,"min_output_million_tokens_price":21.00000000,"max_price":0.0000,"max_input_million_tokens_price":4.00000000,"max_output_million_tokens_price":21.00000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":704,"account":0,"account_type":0,"type":"text2text","service_id":6556,"service_type":"serverless","model_id":25827,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":4.00000000,"output_million_tokens_price":21.00000000,"input_million_tokens_price_average":4.00000000,"output_million_tokens_price_average":21.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":0.8,"cache_write_million_tokens_price":4},"version":1,"specified":{"cache_read_million_tokens_price":0.8,"cache_write_million_tokens_price":4}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1776841017000,"updated_at":1787316785000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[704],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":6699,"ident":"PaddleOCR-VL-1.5","type":"serverless","vendor":"","region":"","compute":"","model":25884,"name":"PaddleOCR-VL-1.5","remark":"PaddleOCR-VL-1.5 是百度 PaddlePaddle 发布的 0.9 B 参数级多任务视觉语言模型，专为复杂文档 OCR 与布局解析设计，在真实世界文档（扫描、倾斜、弯曲等）上实现了先进的识别与结构化输出性能。","price":0.0000,"props":{"name":"PaddleOCR-VL-1.5","service_type":"serverless"},"inventories":0,"created_at":1769771788000,"updated_at":1770013017000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai-testing.su.bcebos.com/v1/uploads/img/id-184hvqui5-avatar-paddle.png","description_html":"\uD83E\uDDE0 模型特点
PaddleOCR-VL-1.5 是百度飞桨推出的下一代 文档解析视觉语言模型（Vision-Language Model），在轻量级（约 0.9 B 参数）下实现业界领先的文档理解能力。
轻量级但性能领先
模型在权威文档解析基准 OmniDocBench v1.5 上取得 94.5 % 的准确率，整体表现优于此前的 SOTA 方案。支持真实场景下的复杂文档解析
通过支持 异形（非矩形）区域检测，可以精准识别倾斜、弯曲、拍照畸变等现实场景下的文本与布局，使 OCR 更适应手机拍摄或扫描件处理。多任务集成能力
除了传统的文本识别外，还具备 文字定位 + 识别（text spotting） 和 印章/印鉴识别（seal recognition） 等任务能力，在各自任务指标上也达到了 SOTA。增强的多语言与特殊元素识别
在多语种环境（如西藏文、孟加拉文等）以及对罕见字符、古籍、表格下划线、复选框等复杂元素的识别能力都有明显提升。长文档与跨页结构处理
支持 跨页表格合并 和 跨页段落标题识别，有效解决长文档拆页导致结构理解断裂的问题。高效推理与部署友好
尽管具有强大多任务能力，模型仍保持紧凑参数规模（0.9 B），便于在服务器或边缘设备上集成部署。
\uD83D\uDCC2 典型使用场景
PaddleOCR-VL-1.5 适用于各种需要精准理解复杂文档结构的实际应用场景：
\uD83D\uDCC4 文档数字化与归档
将扫描件、拍照文档、老旧档案等转换为结构化的文本、JSON 或 Markdown，减少人工校对。\uD83E\uDDFE 票据/合同自动化处理
支持印章识别与复杂布局理解，在财务单据、合同与政务流程自动化中发挥重要作用。\uD83D\uDCCA 科研与技术文档解析
处理含有表格、公式、图表的学术论文、报告与技术文档，准确输出阅读顺序与结构化信息。\uD83D\uDCF1 移动拍照 OCR 应用
针对现实光照、倾斜与弯曲等问题表现稳健，适合手机端扫描与现场数据采集。\uD83C\uDF10 跨语言文档处理
在多语言 OCR 项目中支持更多语种与特殊字符，提高全球化文档理解能力。
DeepSeek-OCR-2 是一款面向真实文档场景优化的高性能 OCR 模型，核心特点包括：
版面理解能力强
不仅能识别文字，还能理解页面结构，自动按照人类阅读顺序输出内容，避免传统 OCR 出现“顺序混乱”的问题。复杂排版识别效果好
在多栏排版、表格、混合图文、票据等复杂版式下，依然能够保持较高的识别准确率和结构完整性。支持多种文档输入形式
可处理扫描图片、拍照图片及 PDF 文档，适用于多种业务场景的文字提取需求。输出结果更易直接使用
识别结果可直接生成结构化文本，便于后续进行搜索、分析、存储或导出为其他格式。适合通过 API 集成到业务系统
可作为 OCR 服务接入各类系统，支持批量文档处理与自动化流程。
使用场景
DeepSeek-OCR-2 API 适用于需要从文档中自动提取文字信息的场景，例如：
文档数字化：
将扫描版合同、报告、论文、档案等转为可编辑文本。表格与票据识别：
自动识别发票、收据、报表、清单中的文字和表格内容。知识库与搜索系统构建：
对大量 PDF 和图片文档进行文字提取，用于全文检索和内容分析。智能办公与自动化处理：
用于审批流、资料录入、报表生成等自动化场景，减少人工录入成本。多栏与复杂版式文档解析：
适合处理杂志、论文、说明书等存在多栏排版和混合结构的页面。
LegalOne-8B 是一款面向中文法律场景优化的 80 亿参数大语言模型，专注于法律文本理解与生成任务。该模型在大规模法律法规、判例和专业语料上训练，具备更强的法律术语理解能力和逻辑推理能力，能够准确处理法律问答、条文解读、案例分析和文书生成等复杂任务，同时在保证专业性的前提下兼顾通用对话能力。
模型特点：
✅ 法律领域专用优化：针对法规条文、司法解释、裁判文书等数据训练，更懂法律语境与表达方式。
✅ 高效中等规模模型：8B 参数量在效果与推理成本之间取得平衡，适合在线 API 调用与规模化部署。
✅ 强推理与长文本理解能力：可处理较长法律文本，实现条款抽取、要点总结和逻辑分析。
✅ 可扩展集成：通过 API 方式接入，便于与现有业务系统、知识库或检索系统结合。
典型使用场景：
⚖ 智能法律咨询：为用户提供法规解释、问题初步分析与合规建议。
\uD83D\uDCC4 法律文书生成与辅助撰写：如合同草稿、起诉状、答辩状、法律意见书初稿等。
\uD83D\uDD0D 案例与法规检索增强：结合搜索系统，对检索结果进行总结、对比和要点提取。
\uD83E\uDDE0 司法与合规辅助分析：用于案件要点归纳、风险点提示与合规检查。
\uD83C\uDFE2 企业法务与政企应用：支持企业内控、合规审查和法务知识问答系统。
\uD83D\uDD0A 模型概述
Qwen3-TTS-12Hz-1.7B-Base 是一个开源的 高质量文本转语音（TTS）生成模型，属于 Qwen3-TTS 系列的一员，由 Qwen 团队发布在 Hugging Face 上。该模型拥有 约1.7 亿参数规模，结合创新的 12 Hz 语音 tokenizer，可实现自然、连贯的多语言语音合成，同时支持 极短语音样本的快速语音克隆。
⭐ 主要特点
\uD83E\uDDE0 1. 3 秒快速语音克隆
只需 约 3 秒的参考音频 即可快速生成目标说话者的语音风格（克隆音色），保留语者的语调和特征，适合个性化语音输出。
\uD83C\uDF0D 多语言支持
支持 10+ 语言 合成，包括汉语、英语、日语、韩语、德语、法语、西班牙语、葡萄牙语等，便于跨语言内容输出。
⚡ 超低延迟、实时流式输出
采用创新的 12 Hz tokenizer 和双轨流式架构，端到端语音合成延迟可低至 ~97 ms，适配实时交互场景，如会话系统或即时翻译。
\uD83C\uDFA4 保留丰富语音信息
模型设计可完整保留旁语信息（如情绪、呼吸和环境声），生成的语音更自然、更富表现力。
\uD83D\uDD04 可用于微调与定制
虽然是 Base 版本，但可以作为基础模型进一步微调（fine-tune）以适配特定语音风格和应用需求。
\uD83D\uDD27 应用场景
\uD83D\uDDE3 1. 个性化语音助手
为聊天机器人、虚拟助手提供自然语音输出，可根据用户或品牌风格克隆独特语音。
\uD83D\uDCDA 2. 有声内容制作
自动生成有声小说、有声读物、播客等内容，特别是在需要快速生成大量语音内容时。
\uD83C\uDF10 3. 跨语言内容转换
支持多种语言合成，可用于国际化平台将文本内容转成不同语种语音，增强全球用户体验。
\uD83D\uDD79 4. 实时交互与游戏
结合实时流式能力，可实现游戏内 NPC 语音对话、互动性教育平台语音反馈等。
\uD83C\uDFA7 5. 辅助工具与创意内容
可集成到工具链中用于自动配音、视频旁白、语音邮件生成或创作 AI 角色声音。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":238,"name":"语音和音频合成","slug":"text-to-speech","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-to-speech.svg","en_name":"Text To Speech","color":"","description":"","href":"","created_at":1700443954770,"sort_order":0,"categories":[{"id":13,"parent_id":1,"name":"音频","slug":"audio","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/audio.svg","sort_order":13,"en_name":"Audio"},{"id":19,"parent_id":8,"name":"音频","slug":"dataset-audio","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-audio.svg","sort_order":19,"en_name":"Audio"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1160,"name":"多语言","slug":"multi-language-label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287760476,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":25778,"name":"Qwen3-TTS-12Hz-1.7B-Base","namespace":"hf-models","mirror_namespace":"Qwen","mirror_url":"https://huggingface.co/Qwen/Qwen3-TTS-12Hz-1.7B-Base"},"total_calls":21,"operation_summary":{"service_id":6853,"unit_tag":1261,"min_price":0.8000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.8000,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":627,"account":0,"account_type":0,"type":"text2speech","service_id":6853,"service_type":"serverless","model_id":25778,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"语音生成(异步)","description":"","path":"v1/async/audio/speech","protocol":"http","price":0.8000,"average_price":0.8000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"output_format","type":"string","control":"select","options":[{"label":"mp3","value":"mp3"}],"required":false,"description":"返回的音频编码方式。","default_value":"mp3"},{"in":"body","name":"inputs","type":"array","control":"array","example":[{"prompt":"其实我真的有发现，我是一个特别善于观察别人情绪的人。","speaker":"Vivian","language":"Chinese","instruction":"体现撒娇稚嫩的萝莉女声，音调偏高且起伏明显，营造出黏人、做作又刻意卖萌的听觉效果。"},{"prompt":"其实我真的有发现，我是一个特别善于观察别人情绪的人。","language":"Chinese","prompt_text":"希望你以后能够做的比我还好哟","prompt_audio_url":"https://gitee.com/realhugh/imgs/raw/master/zero_shot_prompt.wav"}],"samples":[{"label":"示例一","value":[{"prompt":"其实我真的有发现，我是一个特别善于观察别人情绪的人。","speaker":"Vivian","language":"Chinese","instruction":"体现撒娇稚嫩的萝莉女声，音调偏高且起伏明显，营造出黏人、做作又刻意卖萌的听觉效果。"}]},{"label":"示例二","value":[{"prompt":"It's in the top drawer... wait, it's empty? No way, that's impossible! ","speaker":"Vivian","language":"English","instruction":"Speak in an incredulous tone, but with a hint of panic beginning to creep into your voice."}]}],"category":"main","required":true,"max_length":5,"min_length":0,"description":"要转换为语音内容"},{"in":"body","name":"inputs.prompt","type":"string","control":"string","category":"main","required":true,"description":"要转换为语音的文本"},{"in":"body","name":"inputs.language","type":"string","props":{"variant":"select"},"control":"select","options":[{"label":"Chinese","value":"Chinese"},{"label":"English","value":"English"}],"category":"main","required":true,"description":"语言","default_value":"Chinese"},{"in":"body","name":"inputs.speaker","type":"string","props":{"variant":"select"},"control":"select","options":[{"label":"Vivian","value":"Vivian"},{"label":"Serena","value":"Serena"},{"label":"Uncle_Fu","value":"Uncle_Fu"},{"label":"Dylan","value":"Dylan"},{"label":"Eric","value":"Eric"},{"label":"Ryan","value":"Ryan"},{"label":"Aiden","value":"Aiden"},{"label":"Ono_Anna","value":"Ono_Anna"},{"label":"Sohee","value":"Sohee"}],"category":"main","required":false,"description":"音色","default_value":"Ono_Anna"},{"in":"body","name":"inputs.prompt_audio_url","type":"string","control":"string","category":"main","required":false,"description":".wav 格式的文件链接，用于克隆的音色"},{"in":"body","name":"inputs.prompt_text","type":"string","control":"text","category":"main","required":false,"description":"prompt_audio对应的文本"},{"in":"body","name":"inputs.instruction","type":"string","control":"text","category":"main","required":false,"description":"生成语音的指令或风格提示"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1770279739000,"updated_at":1778210622000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[627],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":1261}]},{"service":{"id":6865,"ident":"FLUX.2-klein-4B","type":"serverless","vendor":"","region":"","compute":"","model":25715,"name":"FLUX.2-klein-4B","remark":"FLUX.2-klein-4B 是 Black Forest Labs 发布的一款 轻量级、高速推理的图像生成 AI 模型。","price":0.0000,"props":{"name":"FLUX.2-klein-4B","service_type":"serverless"},"inventories":0,"created_at":1770288190000,"updated_at":1770288274000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai-testing.su.bcebos.com/v1/uploads/img/id-408e8zk0l-avatar-blackforestlab.png","description_html":"模型介绍
FLUX.2-klein-4B 是 Black Forest Labs 推出的轻量级高性能图像生成，拥有约 40 亿参数，基于新一代 rectified flow 架构设计，在保持较高画质的同时大幅提升推理速度。
我们已将该模型封装为标准化 API 服务，用户无需自行部署，即可通过接口快速调用其图像生成能力。
\uD83D\uDE80 模型特点
⚡ 高速推理，适合实时应用
支持极少推理步数即可完成高质量生成，具备接近实时的响应能力。
适用于对延迟敏感的在线应用与交互式场景。
\uD83C\uDFA8 统一的生成与编辑能力
同一模型同时支持：
文本生成图像（Text-to-Image）
可在保持主体一致性的同时完成风格转换与内容调整。
\uD83E\uDDE9 轻量化与高性价比
相比超大规模图像模型，对算力要求更低，推理效率更高。
通过 API 方式调用，可在保证效果的同时降低整体算力成本。
\uD83D\uDCCC 典型应用场景
\uD83D\uDECD 电商与营销素材生成
生成商品展示图、场景图和广告素材。
可基于参考图统一风格，提升品牌视觉一致性。
\uD83D\uDCBB 在线图像生成与编辑应用
适用于网页端、APP、SaaS 工具中的图像生成功能。
通过 API 实现即时预览与交互式修改。
\uD83E\uDDEA AI 图像能力集成
可作为多模态应用中的图像生成模块，与文本、语音系统协同工作。
适合用于智能助手、创作平台和内容工具链。
Z-Image（Tongyi-MAI）模型介绍
Z-Image 是通义实验室推出的高质量图像生成模型，我们已通过 API 方式对外开放，帮助开发者与企业快速获得稳定、可控的文生图能力。
✨ 模型特点
\uD83C\uDFA8 高质量图像生成
具备优秀的细节表现能力，可生成清晰自然、结构合理的高分辨率图像。\uD83E\uDDE0 强语义理解能力
对文本指令具备良好的理解能力，能够准确还原描述中的主体、风格与场景。\uD83E\uDDE9 多风格支持
支持写实、插画、设计稿、概念图等多种视觉风格，适配不同业务需求。\uD83D\uDD0C API 即开即用
无需部署模型环境，通过接口即可调用图像生成能力，快速接入现有系统。⚙ 适合生产级应用
支持批量调用与自动化流程，适用于高并发与规模化业务场景。
\uD83D\uDE80 应用场景
\uD83D\uDCE2 内容创作与营销设计
自动生成海报、宣传图、运营配图等视觉素材。\uD83D\uDED2 电商与商品展示
生成商品场景图、背景图和营销素材，提升展示效果。\uD83C\uDFAE 游戏与数字内容制作
用于角色设定、场景原画和概念设计辅助。\uD83C\uDFEB 教育与科研
用于图像生成研究、创意教学和数据增强。\uD83E\uDDF0 AIGC 平台与工具集成
作为底层图像生成引擎，为创作工具与应用提供能力支持。
\uD83E\uDDE0 模型概述（Model Overview）
Qwen3-Coder-Next 是一款开源权重的因果语言模型，专为编程任务、智能编码代理与本地开发场景设计。它采用 80 B 参数规模（其中约 3 B 激活参数） 的稀疏混合专家（Mixture of Experts，MoE）架构，在计算效率和推理能力之间取得优秀平衡。
\uD83C\uDF1F 核心特点
\uD83D\uDE80 高效能与低运行成本
仅 约 3 B 激活参数 即可实现与传统模型中活跃参数规模达到 10–20× 的性能，推理成本显著更低。
\uD83D\uDEE0 强大的“Agentic”编码能力
针对长流程推理、复杂工具调用和执行错误恢复进行了优化，非常适合用作智能编码代理（coding agents）。
\uD83D\uDCDA 超长上下文支持
原生支持 长达 262 144（约 256 K）令牌的上下文窗口，能处理大规模代码库、文档或多文件工程场景。
\uD83E\uDDE9 灵活集成与多环境适配
与主流开发环境和 CLI/IDE 平台兼容，如 Claude Code、Qwen Code、Qoder、Kilo、Trae、Cline 等，实现便捷集成。
\uD83E\uDDE0 架构创新
使用混合注意力机制与稀疏专家网络（Mixture of Experts）提升模型的推理效率与表达能力。
\uD83D\uDCA1 主要使用场景
\uD83E\uDDD1\uD83D\uDCBB 1. 代码生成与自动补全
可以根据自然语言描述自动生成代码片段、实现函数逻辑、构建模块结构等，大幅提升编码效率。
例如：
自动生成排序算法实现；
根据说明构建 API、数据处理管道。
（可通过 Hugging Face Transformers 接入本地或云 API 实例）
\uD83E\uDDEA 2. 智能调试与错误修复
借助其强大的推理与工具调用能力，模型能辅助识别代码错误、建议修复方案，甚至自动编写测试用例或执行代码片段。
\uD83D\uDEA7 3. 构建智能开发代理（Coding Agents）
适合用于构建可执行外部工具的智能代理，如自动构建脚本、分析大规模代码库、控制版本管理流程等，通过定制工具集可整合到自动化开发链中。
\uD83D\uDCC4 4. 大型文档与多文件工程理解
凭借超长上下文能力，适用于处理大规模 Markdown、代码仓库说明文档、跨文件依赖追踪等任务，使模型在复杂项目理解上更具优势。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":29,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1150,"name":"32K","slug":"context_size_32k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-f8crosrdk-id-y4z8cy4z9-context_size.png","en_name":"32K","color":"blue","description":"上下文大小","href":"","created_at":1734693847615,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":26020,"name":"Qwen3-Coder-Next","namespace":"hf-models","mirror_namespace":"Qwen","mirror_url":"https://huggingface.co/Qwen/Qwen3-Coder-Next"},"total_calls":14,"operation_summary":{"service_id":6900,"unit_tag":0,"min_price":0.0200,"min_input_million_tokens_price":4.00000000,"min_output_million_tokens_price":4.00000000,"max_price":0.0200,"max_input_million_tokens_price":4.00000000,"max_output_million_tokens_price":4.00000000,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":634,"account":0,"account_type":0,"type":"text2text","service_id":6900,"service_type":"serverless","model_id":26020,"vendor":"ascend","region":"prod-as-sz-ip","compute":"ascend/prod-as-sz-ip/serverless_api/gpu/xlarge/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0200,"average_price":0.0200,"input_million_tokens_price":4.00000000,"output_million_tokens_price":4.00000000,"input_million_tokens_price_average":4.00000000,"output_million_tokens_price_average":4.00000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1770370367000,"updated_at":1771845892000,"vendor_info":{"id":0,"name":"ascend","label":"华为昇腾","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-9b8m9pbwi-image 2.svg"},"region_info":{"name":"prod-as-sz-ip","label":"深圳","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"ascend/prod-as-sz-ip/serverless_api/gpu/xlarge/default","instance_size":"xlarge","instance_type":"default","price_per_hour":4.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":192,"memory_gb":1511,"architecture":"Ascend 910","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-as-sz-ip","vendor_name":"ascend","type":"SERVERLESS_API","storage":0},"ids":[634],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":6989,"ident":"GLM-5","type":"serverless","vendor":"","region":"","compute":"","model":26058,"name":"GLM-5","remark":"GLM-5 是面向 Coding 与 Agent 场景的新一代大模型，在复杂系统工程与长程任务中达到开源 SOTA，真实编程体验逼近 Claude Opus 级别；基于 744B 新基座、异步强化学习与稀疏注意力，实现从“写代码”到“写工程”的全面升级。","price":0.0000,"props":{"name":"GLM-5","service_type":"serverless"},"inventories":0,"created_at":1770853836000,"updated_at":1787915247000,"sort_order":0,"status":1,"recommend":true,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-vs79e29q6-avatar-zhipuai.png","description_html":"","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":1,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1262,"name":"200K","slug":"context_size_200k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-v6ul4uqrf-id-9zvdx4rar-id-zd1snxypr-context_size.png","en_name":"200K","color":"blue","description":"上下文大小","href":"","created_at":1759218439298,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1287,"name":"梯度计费","slug":"tiered_pricing","icon":"","en_name":"Tiered Pricing","color":"","description":"","href":"","created_at":1787913527124,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":26058,"name":"GLM-5","namespace":"hf-models","mirror_namespace":"zai-org","mirror_url":"https://huggingface.co/zai-org/GLM-5"},"total_calls":63991,"operation_summary":{"service_id":6989,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":4.00000000,"min_output_million_tokens_price":18.00000000,"max_price":0.0000,"max_input_million_tokens_price":4.00000000,"max_output_million_tokens_price":18.00000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":725,"account":0,"account_type":0,"type":"text2text","service_id":6989,"service_type":"serverless","model_id":26058,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":4.00000000,"output_million_tokens_price":18.00000000,"input_million_tokens_price_average":4.00000000,"output_million_tokens_price_average":18.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"tiered_prompt_cache_tokens","tiers":[{"average":{"input_million_tokens_price":4,"output_million_tokens_price":18,"cache_read_million_tokens_price":0.8},"specified":{"input_million_tokens_price":4,"output_million_tokens_price":18,"cache_read_million_tokens_price":0.8},"max_input_tokens":32000,"min_input_tokens":0},{"average":{"input_million_tokens_price":6,"output_million_tokens_price":22,"cache_read_million_tokens_price":1.2},"specified":{"input_million_tokens_price":6,"output_million_tokens_price":22,"cache_read_million_tokens_price":1.2},"min_input_tokens":32000}],"version":1,"calculation_mode":"whole_order"},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1778752485000,"updated_at":1787914967000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[725],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":7026,"ident":"MiniMax-M2.5","type":"serverless","vendor":"","region":"","compute":"","model":26061,"name":"MiniMax-M2.5","remark":"MiniMax-M2.5 是 MiniMax 编码与智能体领域 SOTA，专为 Agent Universe 而设计，显著增强多语言编程能力，致力于提升真实世界复杂任务中的表现。","price":0.0000,"props":{"name":"MiniMax-M2.5","service_type":"serverless"},"inventories":0,"created_at":1770971999000,"updated_at":1787316751000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-lrzb3ssm1-minimax.webp","description_html":"MiniMax-M2.5 是 MiniMax 编码与智能体领域 SOTA，专为 Agent Universe 而设计，显著增强多语言编程能力，致力于提升真实世界复杂任务中的表现。
模型亮点
相比于前代模型，M2.5 在处理 Agentic 任务时也展现出了更高的决策成熟度：它学会了用更精准的搜索轮次和更优的 token 效率去解决问题。
M2.5 在 Word、PPT、Excel 金融建模等 workspace 高阶场景中取得了显著的能力提升, 在与主流模型的对比中，它取得了 59.0% 的平均胜率。
结合强化学习优化的任务拆解能力与 thinking token 效率，使 M2.5 在完成复杂任务的速度与成本上具备显著优势。
Qwen3.5原生视觉语言系列Flash模型，基于混合架构设计，融合了线性注意力机制与稀疏混合专家模型，实现了更高的推理效率。模型效果在纯文本与多模态方面相较3系列均实现飞跃式进步；响应速度快，兼具推理速度和性能。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1147,"name":"视觉模型","slug":"vision-language","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-vision-language.svg","en_name":"Vision Language","color":"","description":"","href":"","created_at":1734083179939,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1169,"name":"视觉","slug":"vision-language-label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735288027632,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1220,"name":"64K","slug":"max_tokens_64k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-xroamij61-context_size (1).png","en_name":"64K","color":"blue","description":"上下文长度","href":"","created_at":1748699035394,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1287,"name":"梯度计费","slug":"tiered_pricing","icon":"","en_name":"Tiered Pricing","color":"","description":"","href":"","created_at":1787913527124,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":26139,"name":"Qwen3.5-35B-A3B","namespace":"hf-models","mirror_namespace":"Qwen","mirror_url":"https://huggingface.co/Qwen/Qwen3.5-35B-A3B"},"total_calls":117,"operation_summary":{"service_id":7090,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":0.40000000,"min_output_million_tokens_price":3.20000000,"max_price":0.0000,"max_input_million_tokens_price":0.40000000,"max_output_million_tokens_price":3.20000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":696,"account":0,"account_type":0,"type":"text2text","service_id":7090,"service_type":"serverless","model_id":26139,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0.40000000,"output_million_tokens_price":3.20000000,"input_million_tokens_price_average":0.40000000,"output_million_tokens_price_average":3.20000000,"billing_strategy":"rule_based","billing_policy":{"type":"tiered_prompt_cache_tokens","tiers":[{"average":{"input_million_tokens_price":0.4,"output_million_tokens_price":3.2},"specified":{"input_million_tokens_price":0.4,"output_million_tokens_price":3.2},"max_input_tokens":128000,"min_input_tokens":0},{"average":{"input_million_tokens_price":1.6,"output_million_tokens_price":12.8},"specified":{"input_million_tokens_price":1.6,"output_million_tokens_price":12.8},"min_input_tokens":128000}],"version":1,"calculation_mode":"whole_order"},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"最大体积：5MB"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1776416686000,"updated_at":1787914967000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[696],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":7115,"ident":"Qwen3.5-27B","type":"serverless","vendor":"","region":"","compute":"","model":26150,"name":"Qwen3.5-27B","remark":"Qwen3.5 代表了一次重大飞跃，融合了多模态学习、架构效率、强化学习规模以及全球可访问性等方面的突破，为开发者和企业带来前所未有的能力与效率。","price":0.0000,"props":{"name":"Qwen3.5-27B","service_type":"serverless"},"inventories":0,"created_at":1772105290000,"updated_at":1787915183000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3.5 代表了一次重大飞跃，融合了多模态学习、架构效率、强化学习规模以及全球可访问性等方面的突破，为开发者和企业带来前所未有的能力与效率。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1147,"name":"视觉模型","slug":"vision-language","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-vision-language.svg","en_name":"Vision Language","color":"","description":"","href":"","created_at":1734083179939,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1169,"name":"视觉","slug":"vision-language-label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735288027632,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1220,"name":"64K","slug":"max_tokens_64k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-xroamij61-context_size (1).png","en_name":"64K","color":"blue","description":"上下文长度","href":"","created_at":1748699035394,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1287,"name":"梯度计费","slug":"tiered_pricing","icon":"","en_name":"Tiered Pricing","color":"","description":"","href":"","created_at":1787913527124,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":26150,"name":"Qwen3.5-27B","namespace":"hf-models","mirror_namespace":"Qwen","mirror_url":"https://huggingface.co/Qwen/Qwen3.5-27B"},"total_calls":8,"operation_summary":{"service_id":7115,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":0.60000000,"min_output_million_tokens_price":4.80000000,"max_price":0.0000,"max_input_million_tokens_price":0.60000000,"max_output_million_tokens_price":4.80000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":695,"account":0,"account_type":0,"type":"text2text","service_id":7115,"service_type":"serverless","model_id":26150,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0.60000000,"output_million_tokens_price":4.80000000,"input_million_tokens_price_average":0.60000000,"output_million_tokens_price_average":4.80000000,"billing_strategy":"rule_based","billing_policy":{"type":"tiered_prompt_cache_tokens","tiers":[{"average":{"input_million_tokens_price":0.6,"output_million_tokens_price":4.8},"specified":{"input_million_tokens_price":0.6,"output_million_tokens_price":4.8},"max_input_tokens":128000,"min_input_tokens":0},{"average":{"input_million_tokens_price":1.8,"output_million_tokens_price":14.4},"specified":{"input_million_tokens_price":1.8,"output_million_tokens_price":14.4},"min_input_tokens":128000}],"version":1,"calculation_mode":"whole_order"},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"最大体积：5MB"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1776416514000,"updated_at":1787914967000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[695],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":7784,"ident":"Qwen3.5-122B-A10B","type":"serverless","vendor":"","region":"","compute":"","model":26140,"name":"Qwen3.5-122B-A10B","remark":"Qwen3.5-122B-A10B 是一款原生多模态大模型，具备复杂推理、代码生成与智能体能力。","price":0.0000,"props":{"name":"Qwen3.5-122B-A10B","service_type":"serverless"},"inventories":0,"created_at":1773999159000,"updated_at":1787915146000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"\uD83D\uDCCC Qwen3.5-122B-A10B 详细介绍
Qwen3.5-122B-A10B 是一款原生多模态大模型，具备强大的复杂推理、代码生成与智能体能力，能够统一理解与处理文本、图像及视频等多种数据形式，适用于多样化的智能应用场景。
\uD83C\uDF1F 核心特点
\uD83E\uDDE0 强复杂推理能力
支持多步骤逻辑推理与长上下文理解，能够完成复杂问题拆解、链式思考（CoT）以及结构化分析任务，适用于高难度问答与决策辅助。
\uD83D\uDCBB 优秀的代码生成与工程能力
在多语言编程（Java / Python / JS 等）任务中表现突出，支持代码生成、调试、重构与解释，适合开发辅助、自动化工具与AI编程场景。
\uD83D\uDDBC 原生多模态理解
支持文本、图像、视频的统一输入与理解，可完成图像问答、视频内容解析、跨模态推理等任务，无需额外模型拼接。
\uD83E\uDD16 智能体（Agent）能力
具备工具调用与任务规划能力，可作为智能体核心模型，支持复杂任务拆解、多轮交互与自动执行流程。
\uD83D\uDCDA 长上下文处理
支持长文本理解与上下文记忆，适用于文档分析、知识问答、RAG 等场景。
\uD83D\uDE80 应用场景
\uD83D\uDC68\uD83D\uDCBB AI 编程助手 / 开发工具
自动生成代码、接口、SQL、测试用例
代码审查、Bug 定位与优化建议
构建类似 Copilot / Codex / OpenClaw 的开发助手
\uD83D\uDC49 非常适合你的「模力方舟开发者工具链」
\uD83E\uDDE9 智能体与自动化流程（Agent）
自动调用 API / 工具执行任务
多步骤任务规划（如数据分析、报表生成）
构建企业级 AI Agent / Workflow 系统
\uD83D\uDC49 可作为你平台「AI 应用市场」核心能力底座
\uD83D\uDDBC 多模态内容理解与处理
图像问答（识别内容、解释图片）
视频分析（摘要、关键帧理解）
图文结合推理（如电商、审核、教育）
\uD83D\uDC49 可扩展为视觉理解 / 内容审核 / 数据标注能力
\uD83D\uDCCA 企业知识问答 & RAG
企业文档解析（PDF / Wiki / 报告）
知识库问答系统
客服/内部助手
\uD83D\uDC49 非常适合你做「企业AI应用」
✍ 内容生成与智能创作
文案生成、总结、改写
多语言翻译与润色
长内容生成（报告 / 方案 / PRD）
\uD83C\uDFE2 行业智能应用
金融分析与投研辅助
医疗文本理解（非诊断）
教育辅导与解题讲解
GLM-5.1 是智谱最新旗舰模型，代码能力大大增强，长程任务显著提升，能够在单次任务中持续、自主地工作长达 8 小时，完成从规划、执行到迭代优化的完整闭环，交付工程级成果。
在综合能力与 Coding 能力上，GLM-5.1 整体表现对齐 Claude Opus 4.6，并在长程自主执行、复杂工程优化与真实开发场景中展现出更强的持续工作能力，是构建 Autonomous Agent 与长程 Coding Agent 的理想基座。
\uD83E\uDDE0 模型介绍
Gemma 4 26B A4B-it 是一款基于 MoE（Mixture-of-Experts）架构的指令微调大模型，总参数规模为 26B，但单次推理仅激活约 4B 参数，在保证推理能力的同时显著降低计算成本。
该模型针对多步骤推理、结构化输出与工程类任务进行了优化，在性能与成本之间提供稳定平衡，适合高频 API 调用场景。
✨ 核心特点
1⃣ 高效推理架构（MoE）
采用专家混合架构，仅激活部分参数参与计算：
推理开销接近中等规模模型
在性能与资源消耗之间取得平衡
\uD83D\uDC49 特点：
更低计算成本
更高并发适配能力
更稳定的响应表现
2⃣ 稳定的多步骤推理能力
针对复杂任务进行优化：
擅长问题拆解与逐步执行
在多条件、多约束任务中表现稳定
\uD83D\uDC49 特别适合：
需要明确推理过程的任务
规则型与逻辑型场景
3⃣ 结构化输出能力强
对格式类结果支持良好：
JSON 生成
结构化文本输出
信息抽取与字段映射
\uD83D\uDC49 表现：
输出更规整
格式更稳定
更易集成到系统
4⃣ 推理成本可控
在保证能力的同时降低资源消耗：
更适合持续调用
支持规模化部署
\uD83D\uDC49 优势：
成本更可预测
更适合商业化 API
5⃣ 多任务通用能力
除推理外还支持：
文本生成
代码生成与分析
问题分析与总结
\uD83D\uDC49 具备良好的通用性
\uD83C\uDFAF 适用场景
\uD83E\uDDE9 多步骤任务与逻辑推理
复杂问题拆解
规则判断 / 条件推导
任务规划
\uD83D\uDC49 示例：
“设计一个高并发系统架构”
“分析某段业务逻辑执行流程”
\uD83D\uDCBB 编程与工程类任务
代码生成与补全
Bug 分析与修复
接口与系统设计
\uD83D\uDC49 特点：
输出逻辑清晰
更接近工程思维
\uD83D\uDCCA 数据处理与结构化任务
文本 → JSON
信息抽取
数据整理与转换
\uD83D\uDC49 适合：
自动化数据处理
API 返回结构生成
\uD83E\uDD16 AI Agent / 自动化系统
任务拆解（Task Planning）
多步骤执行
工具调用链路
\uD83D\uDC49 适合：
\uD83D\uDC49 Agent + API 编排
\uD83D\uDCDD 内容生成（偏理性）
技术文档
产品说明
分析报告
\uD83D\uDC49 特点：
结构清晰
可读性强
⚠ 使用建议
更适合逻辑型、结构化任务
建议明确输出格式（如 JSON）
对创意写作类任务不是最优选择
MiniMax-M2.7 在编程、工具调用与搜索、办公生产力及其他多个场景下达到或刷新了最新的 SOTA 基准，正式开启了模型递归自我提升（recursive self-improvement）的旅程。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":23,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1262,"name":"200K","slug":"context_size_200k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-v6ul4uqrf-id-9zvdx4rar-id-zd1snxypr-context_size.png","en_name":"200K","color":"blue","description":"上下文大小","href":"","created_at":1759218439298,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":26612,"name":"MiniMax-M2.7","namespace":"hf-models","mirror_namespace":"MiniMaxAI","mirror_url":"https://huggingface.co/MiniMaxAI/MiniMax-M2.7"},"total_calls":422,"operation_summary":{"service_id":8721,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":2.10000000,"min_output_million_tokens_price":8.40000000,"max_price":0.0000,"max_input_million_tokens_price":2.10000000,"max_output_million_tokens_price":8.40000000,"free_operation_count":4,"operation_count":4},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":687,"account":0,"account_type":0,"type":"text2text","service_id":8721,"service_type":"serverless","model_id":26612,"vendor":"infiniai","region":"prod-ia-default-ip","compute":"","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":2.10000000,"output_million_tokens_price":8.40000000,"input_million_tokens_price_average":2.10000000,"output_million_tokens_price_average":8.40000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1776067305000,"updated_at":1786965797000,"vendor_info":{"id":0,"name":"infiniai","label":"无问芯穹","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-sfpgf35pg-news_cover_2-infinigence-avatar.png"},"region_info":{"name":"prod-ia-default-ip","label":"默认","props":{}},"ids":[687],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":693,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":26612,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0},{"id":724,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":26612,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0},{"id":929,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":26612,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["InfiniAI"],"visible":true,"unit_tag":1216},{"id":693,"account":0,"account_type":0,"type":"text2text","service_id":8721,"service_type":"serverless","model_id":26612,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":2.10000000,"output_million_tokens_price":8.40000000,"input_million_tokens_price_average":2.10000000,"output_million_tokens_price_average":8.40000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":0.42,"cache_write_million_tokens_price":2.1},"version":1,"specified":{"cache_read_million_tokens_price":0.42,"cache_write_million_tokens_price":2.1}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1776311404000,"updated_at":1787316545000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[693],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":687,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":26612,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0},{"id":724,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":26612,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0},{"id":929,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":26612,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216},{"id":724,"account":0,"account_type":0,"type":"text2text","service_id":8721,"service_type":"serverless","model_id":26612,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":2.10000000,"output_million_tokens_price":8.40000000,"input_million_tokens_price_average":2.10000000,"output_million_tokens_price_average":8.40000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1778585948000,"updated_at":1787733028000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[724],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":687,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":26612,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0},{"id":693,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":26612,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0},{"id":929,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":26612,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":1216},{"id":929,"account":0,"account_type":0,"type":"text2text","service_id":8721,"service_type":"serverless","model_id":26612,"vendor":"biren-tech","region":"prod-br-sh-wz-ip","compute":"biren-tech/prod-br-sh-wz-ip/serverless_api/gpu/large/contest","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":2.10000000,"output_million_tokens_price":8.40000000,"input_million_tokens_price_average":2.10000000,"output_million_tokens_price_average":8.40000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1787733076000,"updated_at":1787733092000,"vendor_info":{"id":0,"name":"biren-tech","label":"壁仞科技","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-5rdewrmcj-cover-biren.png"},"region_info":{"name":"prod-br-sh-wz-ip","label":"壁仞上海和乌镇","props":{}},"compute_info":{"id":"biren-tech/prod-br-sh-wz-ip/serverless_api/gpu/large/contest","instance_size":"large","instance_type":"contest","price_per_hour":3.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"壁砺系列","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-br-sh-wz-ip","vendor_name":"biren-tech","type":"SERVERLESS_API","storage":0},"ids":[929],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":687,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":26612,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0},{"id":693,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":26612,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0},{"id":724,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":26612,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":1216}]},{"service":{"id":9015,"ident":"Kimi-K2.6","type":"serverless","vendor":"","region":"","compute":"","model":27248,"name":"Kimi-K2.6","remark":"Kimi K2.6 作为国内领先的 Coding 模型，在长程代码任务中的表现取得了突破，面对不同编程语言（如 Rust、Go、Python）和任务场景（如前端、运维、性能优化）均具备更可靠的泛化能力。","price":0.0000,"props":{"name":"Kimi-K2.6","service_type":"serverless"},"inventories":0,"created_at":1776739720000,"updated_at":1787316446000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-eup915vlt-裁剪的圆形图片.png","description_html":"Kimi K2.6 是 Kimi 最新最智能的模型，Kimi K2.6 的通用 Agent、代码、视觉理解等综合能力得到全面提升，其中在博士级难度的完整版人类最后的考试（Humanity’s Last Exam）、在考察模型真实软件工程能力的 SWE-Bench Pro、评估 Agent 深度检索能力的 DeepSearchQA 等基准测试中均取得行业领先的成绩，同时支持文本、图片与视频输入，思考与非思考模式，对话与 Agent 任务。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":22,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1147,"name":"视觉模型","slug":"vision-language","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-vision-language.svg","en_name":"Vision Language","color":"","description":"","href":"","created_at":1734083179939,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1169,"name":"视觉","slug":"vision-language-label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735288027632,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":27248,"name":"Kimi-K2.6","namespace":"hf-models","mirror_namespace":"moonshotai","mirror_url":"https://huggingface.co/moonshotai/Kimi-K2.6"},"total_calls":1459,"operation_summary":{"service_id":9015,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":6.50000000,"min_output_million_tokens_price":27.00000000,"max_price":0.0000,"max_input_million_tokens_price":6.50000000,"max_output_million_tokens_price":27.00000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":703,"account":0,"account_type":0,"type":"text2text","service_id":9015,"service_type":"serverless","model_id":27248,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":6.50000000,"output_million_tokens_price":27.00000000,"input_million_tokens_price_average":6.50000000,"output_million_tokens_price_average":27.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":1.3,"cache_write_million_tokens_price":6.5},"version":1,"specified":{"cache_read_million_tokens_price":1.3,"cache_write_million_tokens_price":6.5}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","type":"number","control":"number","visible":false,"required":false,"description":"temperature","will_be_null":true},{"in":"body","name":"top_p","type":"number","control":"number","visible":false,"required":false,"description":"top_p","will_be_null":true},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","will_be_null":true},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"最大体积：5MB"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1776840649000,"updated_at":1787316473000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[703],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":9050,"ident":"Qwen3.6-Plus","type":"serverless","vendor":"","region":"","compute":"","model":27301,"name":"Qwen3.6-Plus","remark":"Qwen3.6原生视觉语言系列Plus模型，展现出与当前顶尖前沿模型相媲美的卓越性能，模型效果相较3.5系列显著提升。","price":0.0000,"props":{"name":"Qwen3.6-Plus","service_type":"serverless"},"inventories":0,"created_at":1776842472000,"updated_at":1787913963000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3.6原生视觉语言系列Plus模型，展现出与当前顶尖前沿模型相媲美的卓越性能，模型效果相较3.5系列显著提升。模型在Agentic coding、前端编程、Vibe coding等代码能力、多模态万物识别、OCR、物体定位等能力上显著增强。
该模型版本功能等同于快照模型qwen3.6-plus-2026-04-02
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":21,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1147,"name":"视觉模型","slug":"vision-language","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-vision-language.svg","en_name":"Vision Language","color":"","description":"","href":"","created_at":1734083179939,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1169,"name":"视觉","slug":"vision-language-label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735288027632,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1220,"name":"64K","slug":"max_tokens_64k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-xroamij61-context_size (1).png","en_name":"64K","color":"blue","description":"上下文长度","href":"","created_at":1748699035394,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1287,"name":"梯度计费","slug":"tiered_pricing","icon":"","en_name":"Tiered Pricing","color":"","description":"","href":"","created_at":1787913527124,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":27301,"name":"Qwen3.6-Plus","namespace":"fchange1","title":"Qwen3.6-Plus"},"total_calls":3,"operation_summary":{"service_id":9050,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":8.00000000,"min_output_million_tokens_price":48.00000000,"max_price":0.0000,"max_input_million_tokens_price":8.00000000,"max_output_million_tokens_price":48.00000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":707,"account":0,"account_type":0,"type":"text2text","service_id":9050,"service_type":"serverless","model_id":27301,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":8.00000000,"output_million_tokens_price":48.00000000,"input_million_tokens_price_average":8.00000000,"output_million_tokens_price_average":48.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"tiered_prompt_cache_tokens","tiers":[{"average":{"input_million_tokens_price":2,"output_million_tokens_price":12,"cache_read_million_tokens_price":0.2,"cache_write_million_tokens_price":2.5},"specified":{"input_million_tokens_price":2,"output_million_tokens_price":12,"cache_read_million_tokens_price":0.2,"cache_write_million_tokens_price":2.5},"max_input_tokens":128000,"min_input_tokens":0},{"average":{"input_million_tokens_price":8,"output_million_tokens_price":48,"cache_read_million_tokens_price":0.8,"cache_write_million_tokens_price":10},"specified":{"input_million_tokens_price":8,"output_million_tokens_price":48,"cache_read_million_tokens_price":0.8,"cache_write_million_tokens_price":10},"min_input_tokens":128000}],"version":1,"calculation_mode":"whole_order"},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","type":"number","control":"number","visible":false,"required":false,"description":"temperature","will_be_null":true},{"in":"body","name":"top_p","type":"number","control":"number","visible":false,"required":false,"description":"top_p","will_be_null":true},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","will_be_null":true},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"最大体积：5MB"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1776842597000,"updated_at":1787913952000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[707],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":9051,"ident":"Qwen3.6-Max","type":"serverless","vendor":"","region":"","compute":"","model":27302,"name":"Qwen3.6-Max","remark":"Qwen3.6系列中规模最大、综合能力最强的Max模型Preview版本，","price":0.0000,"props":{"name":"Qwen3.6-Max","service_type":"serverless"},"inventories":0,"created_at":1776842962000,"updated_at":1787913815000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3.6系列中规模最大、综合能力最强的Max模型Preview版本，当前开放纯文本模型能力供体验。相较于此前发布的Qwen3-Max和Qwen3.6-Plus，本模型在vibe coding能力上进一步提升、coding agent执行更加高效、前端编程开发能力显著提升；长尾知识能力进一步升级。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","stop_at":1788796800000,"sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":20,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1220,"name":"64K","slug":"max_tokens_64k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-xroamij61-context_size (1).png","en_name":"64K","color":"blue","description":"上下文长度","href":"","created_at":1748699035394,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1287,"name":"梯度计费","slug":"tiered_pricing","icon":"","en_name":"Tiered Pricing","color":"","description":"","href":"","created_at":1787913527124,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":27302,"name":"Qwen3.6-Max","namespace":"fchange1","title":"Qwen3.6-Max"},"total_calls":8,"operation_summary":{"service_id":9051,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":9.00000000,"min_output_million_tokens_price":54.00000000,"max_price":0.0000,"max_input_million_tokens_price":9.00000000,"max_output_million_tokens_price":54.00000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":708,"account":0,"account_type":0,"type":"text2text","service_id":9051,"service_type":"serverless","model_id":27302,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":9.00000000,"output_million_tokens_price":54.00000000,"input_million_tokens_price_average":9.00000000,"output_million_tokens_price_average":54.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"tiered_prompt_cache_tokens","tiers":[{"average":{"input_million_tokens_price":9,"output_million_tokens_price":54,"cache_read_million_tokens_price":0.9,"cache_write_million_tokens_price":11.25},"specified":{"input_million_tokens_price":9,"output_million_tokens_price":54,"cache_read_million_tokens_price":0.9,"cache_write_million_tokens_price":11.25},"max_input_tokens":128000,"min_input_tokens":0},{"average":{"input_million_tokens_price":15,"output_million_tokens_price":90,"cache_read_million_tokens_price":1.5,"cache_write_million_tokens_price":18.75},"specified":{"input_million_tokens_price":15,"output_million_tokens_price":90,"cache_read_million_tokens_price":1.5,"cache_write_million_tokens_price":18.75},"min_input_tokens":128000}],"version":1,"calculation_mode":"whole_order"},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1776843043000,"updated_at":1787913802000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[708],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":9138,"ident":"DeepSeek-V4-Pro","type":"serverless","vendor":"","region":"","compute":"","model":27395,"name":"DeepSeek-V4-Pro","remark":"旗舰级 MoE 大模型，总参1.6T、激活 49B，原生支持百万级超长上下文。依托海量高质量训练数据，具备顶尖数学逻辑、复杂推理、专业代码与长文本深度解析能力，适配高阶科研、复杂办公、深度智能代理等高难度场景。","price":0.0000,"props":{"name":"DeepSeek-V4-Pro","service_type":"serverless"},"inventories":0,"created_at":1777027997000,"updated_at":1787316294000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-00z769y98-deepseek.png","description_html":"旗舰级 MoE 大模型，总参1.6T、激活 49B，原生支持百万级超长上下文。依托海量高质量训练数据，具备顶尖数学逻辑、复杂推理、专业代码与长文本深度解析能力，适配高阶科研、复杂办公、深度智能代理等高难度场景。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":19,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":27395,"name":"DeepSeek-V4-Pro","namespace":"hf-models","mirror_namespace":"deepseek-ai","mirror_url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro"},"total_calls":1145,"operation_summary":{"service_id":9138,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":12.00000000,"min_output_million_tokens_price":24.00000000,"max_price":0.0000,"max_input_million_tokens_price":12.00000000,"max_output_million_tokens_price":24.00000000,"free_operation_count":2,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":712,"account":0,"account_type":0,"type":"text2text","service_id":9138,"service_type":"serverless","model_id":27395,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":12.00000000,"output_million_tokens_price":24.00000000,"input_million_tokens_price_average":12.00000000,"output_million_tokens_price_average":24.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":1,"cache_write_million_tokens_price":12},"version":1,"specified":{"cache_read_million_tokens_price":1,"cache_write_million_tokens_price":12}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1777028121000,"updated_at":1787554782000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[712],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":785,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":27395,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216},{"id":785,"account":0,"account_type":0,"type":"text2text","service_id":9138,"service_type":"serverless","model_id":27395,"vendor":"ztzn","region":"ztzn-region","compute":"ztzn/ztzn-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":12.00000000,"output_million_tokens_price":24.00000000,"input_million_tokens_price_average":12.00000000,"output_million_tokens_price_average":24.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":1,"cache_write_million_tokens_price":12},"version":1,"specified":{"cache_read_million_tokens_price":1,"cache_write_million_tokens_price":12}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1782898992000,"updated_at":1787556394000,"vendor_info":{"id":0,"name":"ztzn","label":"正拓智能","status":"available","ratio":0.70},"region_info":{"name":"ztzn-region","label":"正拓智能默认","props":{}},"compute_info":{"id":"ztzn/ztzn-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"正拓智能","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"ztzn-region","vendor_name":"ztzn","type":"SERVERLESS_API","storage":0},"ids":[785],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":712,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":27395,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":9139,"ident":"DeepSeek-V4-Flash","type":"serverless","vendor":"","region":"","compute":"","model":27396,"name":"DeepSeek-V4-Flash","remark":"高效轻量化MoE模型，总参284B，激活13B，原生支持百万超长上下文能力。推理速度快、延迟低、调用成本低廉，综合能力均衡，主打高并发、轻量化任务，适合日常对话、内容创作、基础 RAG、批量文案处理等普惠刚需场景。","price":0.0000,"props":{"name":"DeepSeek-V4-Flash","service_type":"serverless"},"inventories":0,"created_at":1777028607000,"updated_at":1787316207000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-00z769y98-deepseek.png","description_html":"高效轻量化MoE模型，总参284B，激活13B，原生支持百万超长上下文能力。推理速度快、延迟低、调用成本低廉，综合能力均衡，主打高并发、轻量化任务，适合日常对话、内容创作、基础 RAG、批量文案处理等普惠刚需场景。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":18,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":27396,"name":"DeepSeek-V4-Flash","namespace":"hf-models","mirror_namespace":"deepseek-ai","mirror_url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash"},"total_calls":535,"operation_summary":{"service_id":9139,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":1.00000000,"min_output_million_tokens_price":2.00000000,"max_price":0.0000,"max_input_million_tokens_price":1.00000000,"max_output_million_tokens_price":2.00000000,"free_operation_count":2,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":713,"account":0,"account_type":0,"type":"text2text","service_id":9139,"service_type":"serverless","model_id":27396,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":1.00000000,"output_million_tokens_price":2.00000000,"input_million_tokens_price_average":1.00000000,"output_million_tokens_price_average":2.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":0.2,"cache_write_million_tokens_price":1},"version":1,"specified":{"cache_read_million_tokens_price":0.2,"cache_write_million_tokens_price":1}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1777028701000,"updated_at":1787554803000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[713],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":786,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":27396,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216},{"id":786,"account":0,"account_type":0,"type":"text2text","service_id":9139,"service_type":"serverless","model_id":27396,"vendor":"ztzn","region":"ztzn-region","compute":"ztzn/ztzn-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":1.00000000,"output_million_tokens_price":2.00000000,"input_million_tokens_price_average":1.00000000,"output_million_tokens_price_average":2.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":0.2,"cache_write_million_tokens_price":1},"version":1,"specified":{"cache_read_million_tokens_price":0.2,"cache_write_million_tokens_price":1}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1782899129000,"updated_at":1788320533000,"vendor_info":{"id":0,"name":"ztzn","label":"正拓智能","status":"available","ratio":0.70},"region_info":{"name":"ztzn-region","label":"正拓智能默认","props":{}},"compute_info":{"id":"ztzn/ztzn-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"正拓智能","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"ztzn-region","vendor_name":"ztzn","type":"SERVERLESS_API","storage":0},"ids":[786],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":713,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":27396,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":9228,"ident":"Qwen3.6-35B-A3B","type":"serverless","vendor":"","region":"","compute":"","model":27529,"name":"Qwen3.6-35B-A3B","remark":"Qwen3.6-35B-A3B 模型是一款基于混合专家架构的高性能大语言模型，具备强推理能力与高性价比，适用于复杂任务与高并发AI应用场景。","price":0.0000,"props":{"name":"Qwen3.6-35B-A3B","service_type":"serverless"},"inventories":0,"created_at":1777376132000,"updated_at":1777536994000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3.6-35B-A3B 是一款基于混合专家架构（MoE）的新一代大语言模型，在保持强大推理能力的同时显著优化计算成本与响应效率。模型在激活参数规模与整体性能之间取得良好平衡，既能处理复杂逻辑推理与多轮对话任务，又具备较高的吞吐能力，适合生产环境部署。
该模型在多语言理解、指令执行、代码生成及复杂任务处理方面表现优异，能够满足从通用文本生成到专业应用场景的多样需求。通过标准 API 接入，开发者可以快速将其集成到各类应用中，实现智能化升级。
\uD83D\uDE80 核心能力
\uD83E\uDDE0 复杂推理能力：支持多步骤推理、逻辑分析与问题拆解
\uD83D\uDCAC 多轮对话理解：上下文一致性强，适用于长对话场景
✍ 高质量文本生成：内容创作、改写、总结、扩写
\uD83D\uDCBB 代码能力：支持代码生成、优化与调试辅助
\uD83C\uDF0D 多语言支持：中英文及多语种处理能力优秀
\uD83D\uDCCA 结构化处理：信息抽取、分类、格式化输出
\uD83C\uDFAF 应用场景
\uD83E\uDD16 智能客服 / AI助手：支持复杂问答与上下文理解，提升用户交互体验
✍ 内容生产：营销文案、SEO文章、社媒内容批量生成
\uD83D\uDCBB 编程助手：代码生成、问题排查、开发提效
\uD83D\uDCCA 企业数据处理：文本分类、信息抽取、自动化流程处理
\uD83E\uDDFE 知识库问答（RAG）：结合企业数据实现精准问答
\uD83C\uDF0D 翻译与本地化：多语言内容生成与转换
⭐ 模型特色
⚡ MoE 架构（高性价比）：以较低计算成本实现接近超大模型的效果
\uD83E\uDDE9 强推理能力：适用于复杂业务逻辑与高难度任务
\uD83D\uDE80 高吞吐 & 低延迟：适合 API 高并发调用场景
\uD83D\uDCCF 长上下文支持：适合长文本处理与复杂对话
\uD83D\uDD27 工程友好：易于通过 OpenAI 兼容接口快速接入
\uD83D\uDCCC 模型简介
Qwen3.6-27B 是一款面向生产环境优化的通用大模型，在推理能力、生成质量与性能之间实现良好平衡。相比更大规模模型，该模型在保证较强理解与生成能力的同时，具备更优的响应速度与资源消耗表现，适合高并发API服务场景。
模型在多语言理解、指令执行、多轮对话以及代码生成等方面表现稳定，能够覆盖从基础文本处理到复杂业务逻辑的多种需求。通过标准化API接口接入，可快速集成到各类应用系统中，降低AI能力落地成本。
\uD83D\uDE80 核心能力
\uD83E\uDDE0 通用推理能力：支持逻辑分析、问题拆解与复杂任务处理
\uD83D\uDCAC 多轮对话：上下文理解稳定，适用于连续交互场景
✍ 文本生成与处理：生成、改写、总结、扩写等
\uD83D\uDCBB 代码能力：支持代码生成、解释与基础调试
\uD83C\uDF0D 多语言支持：中英文及多语种处理能力良好
\uD83D\uDCCA 结构化输出：信息抽取、分类与格式化能力
\uD83C\uDFAF 应用场景
\uD83E\uDD16 AI客服 / 智能助手：支持业务问答、自动回复与用户引导
✍ 内容生成：营销文案、文章写作、社媒内容批量生成
\uD83D\uDCBB 开发辅助：代码生成、接口文档生成、问题排查
\uD83D\uDCCA 数据处理自动化：文本分类、信息抽取、流程自动化
\uD83E\uDDFE 知识库问答（RAG）：结合企业数据实现精准问答
\uD83C\uDF0D 翻译与本地化：多语言内容生成与转换
⭐ 模型特色
⚖ 性能与成本平衡：相比超大模型更具性价比
⚡ 响应速度快：适合在线服务与高并发调用
\uD83E\uDDE9 能力全面：覆盖文本、对话、代码等多类任务
\uD83D\uDD27 工程友好：支持OpenAI兼容接口，易于集成
\uD83D\uDCC8 稳定性强：适合长期运行的生产环境
HappyHorse-1.0-I2V支持图生视频，具备高度还原的动态画面生成能力，能够精准理解文本语义，输出流畅自然、细节丰富的高质量视频。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1191,"name":"视频生成","slug":"video-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text_to_video.svg","en_name":"Video Generation","color":"","description":"","href":"","created_at":1739973857876,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":27923,"name":"HappyHorse-1.0","namespace":"fchange1","title":"HappyHorse-1.0"},"total_calls":3,"operation_summary":{"service_id":9596,"unit_tag":1264,"min_price":0.0010,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0010,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":4},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{"720P":{"name":"720P","width":1280,"height":720,"price_factor":720.0,"pixels":921600},"1080P":{"name":"1080P","width":1920,"height":1080,"price_factor":1280.0,"pixels":2073600}},"all_vendors":false,"subscription_service":false},"operations":[{"id":716,"account":0,"account_type":0,"type":"text2video","service_id":9596,"service_type":"serverless","model_id":27923,"vendor":"aliyun","region":"dashscope","compute":"dashscope","name":"文生视频","description":"HappyHorse 1.0 文生视频","path":"v1/async/videos/generations","protocol":"http","price":0.0010,"average_price":0.0010,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":10,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"720P"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":15,"minimum":3,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"body","expr":"$.ratio","name":"ratio","type":"string","control":"select","options":[{"label":"16:9","value":"16:9"},{"label":"9:16","value":"9:16"},{"label":"1:1","value":"1:1"}],"required":false,"description":"画面比例","default_value":"16:9"},{"in":"body","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机种子"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1778134319000,"updated_at":1782122990000,"ids":[716],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264},{"id":721,"account":0,"account_type":0,"type":"text2video","service_id":9596,"service_type":"serverless","model_id":27923,"vendor":"aliyun","region":"dashscope","compute":"dashscope","name":"参考图视频","description":"HappyHorse 1.0 参考图视频","path":"v1/async/videos/generations","protocol":"http","price":0.0010,"average_price":0.0010,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":20,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"720P"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":15,"minimum":3,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"body","expr":"$.ratio","name":"ratio","type":"string","control":"select","options":[{"label":"16:9","value":"16:9"},{"label":"9:16","value":"9:16"},{"label":"1:1","value":"1:1"}],"required":false,"description":"画面比例","default_value":"16:9"},{"in":"body","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机种子"},{"in":"body","name":"reference_images","type":"array","props":{"textField":false,"imageField":{"name":"","accept":"image/*","max_length":9,"show_array":true}},"accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"reference_image"}},"control":"multimodal","category":"main","encoding":"data-url","required":true,"max_length":9,"min_length":1,"description":"参考图片"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1778134319000,"updated_at":1782122990000,"ids":[721],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264},{"id":720,"account":0,"account_type":0,"type":"text2video","service_id":9596,"service_type":"serverless","model_id":27923,"vendor":"aliyun","region":"dashscope","compute":"dashscope","name":"图生视频","description":"HappyHorse 1.0 首帧图生视频","path":"v1/async/videos/generations","protocol":"http","price":0.0010,"average_price":0.0010,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":30,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"720P"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":15,"minimum":3,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"body","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机种子"},{"in":"body","expr":"$.first_frame","name":"first_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"first_frame"}},"control":"file","category":"main","encoding":"data-url","required":true,"description":"首帧图片"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1778134319000,"updated_at":1782122990000,"ids":[720],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264},{"id":719,"account":0,"account_type":0,"type":"image_video2video","service_id":9596,"service_type":"serverless","model_id":27923,"vendor":"aliyun","region":"dashscope","compute":"dashscope","name":"视频编辑","description":"HappyHorse 1.0 视频编辑","path":"v1/async/videos/image-video-to-video","protocol":"http","price":0.0010,"average_price":0.0010,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":40,"params":[{"in":"form","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":false,"description":"视频提示词"},{"in":"form","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"720P"},{"in":"form","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":15,"minimum":3,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"form","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机种子"},{"in":"form","expr":"$.video","name":"video","type":"file","accept":"video/*","adapter":{"aliyun":{"emit":"media","media_type":"video"}},"control":"file","category":"main","required":true,"description":"待编辑视频，支持上传文件或视频 URL"},{"in":"form","name":"reference_images","type":"array","items":{"type":"file"},"accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"reference_image"}},"control":"file","category":"main","required":false,"max_length":9,"description":"参考图片，可选，最多9张"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1778134319000,"updated_at":1782122990000,"ids":[719],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264}]},{"service":{"id":9951,"ident":"Wan2.7","type":"serverless","vendor":"","region":"","compute":"","model":28036,"name":"Wan 2.7 - 万相视频生成模型","remark":"万相2.7，演绎能力全面升级，文戏情感细腻自然，动作戏激烈拳拳到肉，搭配更富有戏剧性和节奏感的镜头切换，实现更强表演能力。","price":0.0000,"props":{"name":"Wan-2.7","service_type":"serverless","resolution_price_factors":{"720P":{"width":1280,"factor":6.0,"height":720},"1080P":{"width":1920,"factor":10.0,"height":1080}}},"inventories":0,"created_at":1778828897000,"updated_at":1779246326000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"万相2.7，演绎能力全面升级，文戏情感细腻自然，动作戏激烈拳拳到肉，搭配更富有戏剧性和节奏感的镜头切换，实现更强表演能力。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1191,"name":"视频生成","slug":"video-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text_to_video.svg","en_name":"Video Generation","color":"","description":"","href":"","created_at":1739973857876,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":28036,"name":"Wan-2.7","namespace":"fchange1","title":"Wan-2.7"},"total_calls":19,"operation_summary":{"service_id":9951,"unit_tag":1264,"min_price":0.1000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.1000,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":4},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{"720P":{"name":"720P","width":1280,"height":720,"price_factor":6.0,"pixels":921600},"1080P":{"name":"1080P","width":1920,"height":1080,"price_factor":10.0,"pixels":2073600}},"all_vendors":false,"subscription_service":false},"operations":[{"id":727,"account":0,"account_type":0,"type":"text2video","service_id":9951,"service_type":"serverless","model_id":28036,"vendor":"aliyun","region":"dashscope","compute":"dashscope","name":"文生视频","description":"Wan 2.7 文生视频","path":"v1/async/videos/generations","protocol":"http","price":0.1000,"average_price":0.1000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":10,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.negative_prompt","name":"negative_prompt","type":"string","control":"text","required":false,"description":"反向提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"720P"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":15,"minimum":3,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"body","expr":"$.ratio","name":"ratio","type":"string","control":"select","options":[{"label":"16:9","value":"16:9"},{"label":"9:16","value":"9:16"},{"label":"1:1","value":"1:1"}],"required":false,"description":"画面比例","default_value":"16:9"},{"in":"body","expr":"$.prompt_extend","name":"prompt_extend","type":"boolean","control":"boolean","required":false,"description":"是否开启提示词改写","default_value":true},{"in":"body","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机种子"},{"in":"body","expr":"$.audio_url","name":"audio_url","type":"string","accept":"audio/*","control":"file","category":"main","encoding":"data-url","required":false,"description":"驱动音频，支持上传文件或音频 URL"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1778644318000,"updated_at":1778815602000,"ids":[727],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264},{"id":728,"account":0,"account_type":0,"type":"text2video","service_id":9951,"service_type":"serverless","model_id":28036,"vendor":"aliyun","region":"dashscope","compute":"dashscope","name":"参考生视频","description":"Wan 2.7 参考生视频","path":"v1/async/videos/generations","protocol":"http","price":0.1000,"average_price":0.1000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":20,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.negative_prompt","name":"negative_prompt","type":"string","control":"text","required":false,"description":"反向提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"720P"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":15,"minimum":3,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"body","expr":"$.ratio","name":"ratio","type":"string","control":"select","options":[{"label":"16:9","value":"16:9"},{"label":"9:16","value":"9:16"},{"label":"1:1","value":"1:1"}],"required":false,"description":"画面比例","default_value":"16:9"},{"in":"body","expr":"$.prompt_extend","name":"prompt_extend","type":"boolean","control":"boolean","required":false,"description":"是否开启提示词改写","default_value":true},{"in":"body","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机种子"},{"in":"body","name":"reference_images","type":"array","props":{"textField":false,"imageField":{"name":"","accept":"image/*","max_length":9,"show_array":true}},"accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"reference_image"}},"control":"multimodal","category":"main","encoding":"data-url","required":true,"max_length":9,"min_length":1,"description":"参考图片"},{"in":"body","name":"reference_videos","type":"array","items":{"type":"file"},"accept":"video/*","adapter":{"aliyun":{"emit":"media","media_type":"reference_video"}},"control":"file","category":"main","required":false,"max_length":9,"description":"参考视频，可选，最多9个"},{"in":"body","expr":"$.first_frame","name":"first_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"first_frame"}},"control":"file","category":"main","encoding":"data-url","required":false,"description":"首帧图片，可选"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1778644318000,"updated_at":1778815602000,"ids":[728],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264},{"id":726,"account":0,"account_type":0,"type":"text2video","service_id":9951,"service_type":"serverless","model_id":28036,"vendor":"aliyun","region":"dashscope","compute":"dashscope","name":"图生视频","description":"Wan 2.7 通用图生视频","path":"v1/async/videos/generations","protocol":"http","price":0.1000,"average_price":0.1000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":30,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.negative_prompt","name":"negative_prompt","type":"string","control":"text","required":false,"description":"反向提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"720P"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":15,"minimum":3,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"body","expr":"$.ratio","name":"ratio","type":"string","control":"select","options":[{"label":"16:9","value":"16:9"},{"label":"9:16","value":"9:16"},{"label":"1:1","value":"1:1"}],"required":false,"description":"画面比例","default_value":"16:9"},{"in":"body","expr":"$.prompt_extend","name":"prompt_extend","type":"boolean","control":"boolean","required":false,"description":"是否开启提示词改写","default_value":true},{"in":"body","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机种子"},{"in":"body","expr":"$.first_frame","name":"first_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"first_frame"}},"control":"file","category":"main","encoding":"data-url","required":true,"description":"首帧图片"},{"in":"body","expr":"$.last_frame","name":"last_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"last_frame"}},"control":"file","category":"main","encoding":"data-url","required":false,"description":"尾帧图片"},{"in":"body","expr":"$.driving_audio","name":"driving_audio","type":"file","accept":"audio/*","adapter":{"aliyun":{"emit":"media","media_type":"driving_audio"}},"control":"file","category":"main","encoding":"data-url","required":false,"description":"驱动音频"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1778641431000,"updated_at":1778815391000,"ids":[726],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264},{"id":729,"account":0,"account_type":0,"type":"image_video2video","service_id":9951,"service_type":"serverless","model_id":28036,"vendor":"aliyun","region":"dashscope","compute":"dashscope","name":"视频编辑","description":"Wan 2.7 视频编辑","path":"v1/async/videos/image-video-to-video","protocol":"http","price":0.1000,"average_price":0.1000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":40,"params":[{"in":"form","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":false,"description":"视频提示词"},{"in":"form","expr":"$.negative_prompt","name":"negative_prompt","type":"string","control":"text","required":false,"description":"反向提示词"},{"in":"form","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"720P"},{"in":"form","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":15,"minimum":3,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"form","expr":"$.ratio","name":"ratio","type":"string","control":"select","options":[{"label":"16:9","value":"16:9"},{"label":"9:16","value":"9:16"},{"label":"1:1","value":"1:1"}],"required":false,"description":"画面比例","default_value":"16:9"},{"in":"form","expr":"$.prompt_extend","name":"prompt_extend","type":"boolean","control":"boolean","required":false,"description":"是否开启提示词改写","default_value":true},{"in":"form","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机种子"},{"in":"form","expr":"$.video","name":"video","type":"file","accept":"video/*","adapter":{"aliyun":{"emit":"media","media_type":"video"}},"control":"file","category":"main","required":true,"description":"待编辑视频，支持上传文件或视频 URL"},{"in":"form","name":"reference_images","type":"array","items":{"type":"file"},"accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"reference_image"}},"control":"file","category":"main","required":false,"max_length":9,"description":"参考图片，可选，最多9张"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1778644318000,"updated_at":1778815391000,"ids":[729],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264}]},{"service":{"id":10741,"ident":"ViduQ3-Turbo","type":"serverless","vendor":"","region":"","compute":"","model":28207,"name":"ViduQ3-Turbo","remark":"Vidu 的 ViduQ3-Turbo 是一款高性能加速版视频生成模型，兼顾生成效率、画质表现与动态效果，尤其在打斗动作、情绪演绎和语义理解方面表现突出，具备较高性价比，适用于图片社交、AI陪伴、特效素材等泛娱乐场景。","price":0.0000,"props":{"name":"viduq3-turbo","service_type":"serverless","resolution_price_factors":{"540P":{"width":1024,"factor":2.5,"height":576},"720P":{"width":1280,"factor":3.75,"height":720},"1080P":{"width":1920,"factor":4.375,"height":1080}}},"inventories":0,"created_at":1779276050000,"updated_at":1779276050000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xaxydsli0-O1CN01TZi8tC24JyWxZuYyj_!!6000000007371-2-tps-84-84.png","description_html":"","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1191,"name":"视频生成","slug":"video-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text_to_video.svg","en_name":"Video Generation","color":"","description":"","href":"","created_at":1739973857876,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":28207,"name":"viduq3-turbo","namespace":"XuJingZhu","title":"viduq3-turbo"},"total_calls":4,"operation_summary":{"service_id":10741,"unit_tag":1264,"min_price":0.1000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.1000,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":3},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{"540P":{"name":"540P","width":1024,"height":576,"price_factor":2.5,"pixels":589824},"720P":{"name":"720P","width":1280,"height":720,"price_factor":3.75,"pixels":921600},"1080P":{"name":"1080P","width":1920,"height":1080,"price_factor":4.375,"pixels":2073600}},"all_vendors":false,"subscription_service":false},"operations":[{"id":735,"account":0,"account_type":0,"type":"text2video","service_id":10741,"service_type":"serverless","model_id":28207,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文生视频","description":"","path":"v1/async/videos/generations","protocol":"http","price":0.1000,"average_price":0.1000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"540P","value":"540P"},{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"720P"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":16,"minimum":5,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"body","expr":"$.audio","name":"audio","type":"boolean","control":"boolean","required":false,"description":"是否生成有声视频。开启后模型将根据视频内容自动生成匹配的背景音乐或音效。","default_value":false},{"in":"body","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机数种子，取值范围为[0, 2147483647]。"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1779276251000,"updated_at":1780033963000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[735],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264},{"id":736,"account":0,"account_type":0,"type":"text2video","service_id":10741,"service_type":"serverless","model_id":28207,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"基于首尾帧","description":"输入首帧图片、尾帧图片与文本描述，生成视频。","path":"v1/async/videos/generations","protocol":"http","price":0.1000,"average_price":0.1000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"540P","value":"540P"},{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"720P"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":16,"minimum":5,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"body","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机数种子，取值范围为[0, 2147483647]。"},{"in":"body","expr":"$.first_frame","name":"first_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"image"}},"control":"file","category":"main","encoding":"data-url","required":true,"description":"首帧图片"},{"in":"body","expr":"$.last_frame","name":"last_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"image"}},"control":"file","category":"main","encoding":"data-url","required":true,"description":"尾帧图片"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1779276354000,"updated_at":1779448820000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[736],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264},{"id":745,"account":0,"account_type":0,"type":"text2video","service_id":10741,"service_type":"serverless","model_id":28207,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"图生视频","description":"","path":"v1/async/videos/generations","protocol":"http","price":0.1000,"average_price":0.1000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"540P","value":"540P"},{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"720P"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":16,"minimum":5,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"form","expr":"$.watermark","name":"watermark","type":"boolean","control":"boolean","required":false,"description":"是否添加水印标识，水印位于视频右下角，文案固定为“内容由AI生成”。","default_value":false},{"in":"form","expr":"$.audio","name":"audio","type":"boolean","control":"boolean","required":false,"description":"是否生成有声视频。开启后模型将根据视频内容自动生成匹配的背景音乐或音效。","default_value":false},{"in":"body","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机种子"},{"in":"body","expr":"$.first_frame","name":"first_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"image"}},"control":"file","category":"main","encoding":"data-url","required":true,"description":"首帧图片"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1779353567000,"updated_at":1779354162000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[745],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264}]},{"service":{"id":10745,"ident":"ViduQ2-Turbo","type":"serverless","vendor":"","region":"","compute":"","model":28211,"name":"ViduQ2-Turbo","remark":"ViduQ2-Turbo图生视频是极速生成引擎。720P 5s视频最快仅需19秒，1080P 5s视频约27秒。人物动作与表情自然逼真，真实感强，在打斗等高动态场景中效果出色，运动幅度大。","price":0.0000,"props":{"name":"ViduQ2-Turbo","service_type":"serverless","resolution_price_factors":{"540P":{"width":960,"factor":875,"height":540},"720P":{"width":1280,"factor":2500,"height":720},"1080P":{"width":1920,"factor":4688,"height":1080}}},"inventories":0,"created_at":1779277906000,"updated_at":1779277906000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai-testing.su.bcebos.com/v1/uploads/img/id-ypi74repw-20260520-195004.jpg","description_html":"","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1191,"name":"视频生成","slug":"video-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text_to_video.svg","en_name":"Video Generation","color":"","description":"","href":"","created_at":1739973857876,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":28211,"name":"ViduQ2-Turbo","namespace":"moark","title":"ViduQ2-Turbo"},"total_calls":5,"operation_summary":{"service_id":10745,"unit_tag":1264,"min_price":0.0001,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0001,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":3},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{"540P":{"name":"540P","width":960,"height":540,"price_factor":875.0,"pixels":518400},"720P":{"name":"720P","width":1280,"height":720,"price_factor":2500.0,"pixels":921600},"1080P":{"name":"1080P","width":1920,"height":1080,"price_factor":4688.0,"pixels":2073600}},"all_vendors":false,"subscription_service":false},"operations":[{"id":749,"account":0,"account_type":0,"type":"text2video","service_id":10745,"service_type":"serverless","model_id":28211,"vendor":"aliyun","region":"dashscope","compute":"dashscope","name":"文生视频","description":"输入文本描述，生成视频。","path":"v1/async/videos/generations","protocol":"http","price":0.0001,"average_price":0.0001,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":10,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"540P","value":"540P"},{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"720P"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":10,"minimum":5,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"body","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机数种子，取值范围为[0, 2147483647]。"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1780034377000,"updated_at":1780034519000,"ids":[749],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264},{"id":737,"account":0,"account_type":0,"type":"text2video","service_id":10745,"service_type":"serverless","model_id":28211,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"基于首尾帧","description":"输入首帧图片、尾帧图片与文本描述，生成视频。","path":"v1/async/videos/generations","protocol":"http","price":0.0001,"average_price":0.0001,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":20,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"540P","value":"540P"},{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"720P"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":10,"minimum":5,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"body","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机数种子，取值范围为[0, 2147483647]。"},{"in":"body","expr":"$.first_frame","name":"first_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"image"}},"control":"file","category":"main","encoding":"data-url","required":true,"description":"首帧图片"},{"in":"body","expr":"$.last_frame","name":"last_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"image"}},"control":"file","category":"main","encoding":"data-url","required":true,"description":"尾帧图片"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1779277906000,"updated_at":1779448820000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[737],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264},{"id":738,"account":0,"account_type":0,"type":"text2video","service_id":10745,"service_type":"serverless","model_id":28211,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"图生视频","description":"输入首帧图片与文本描述，生成视频。","path":"v1/async/videos/generations","protocol":"http","price":0.0001,"average_price":0.0001,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":30,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"540P","value":"540P"},{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"540P"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":16,"minimum":5,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"body","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机种子"},{"in":"body","expr":"$.first_frame","name":"first_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"image"}},"control":"file","category":"main","encoding":"data-url","required":true,"description":"首帧图片"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1779277906000,"updated_at":1779347603000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[738],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264}]},{"service":{"id":10746,"ident":"ViduQ2-Pro","type":"serverless","vendor":"","region":"","compute":"","model":28210,"name":"ViduQ2-Pro","remark":"ViduQ2-Pro图生视频是全球首创「万物可参考」视频模型。支持特效、表情、纹理、动作、人物、场景等六大维度参考，实现编辑全面进化。通过可控式增、删、改，达成精细化视频编辑，专为漫剧、短剧、影视制作打造的生产级创作引擎","price":0.0000,"props":{"name":"ViduQ2-Pro","service_type":"serverless","resolution_price_factors":{"540P":{"width":960,"factor":1,"height":540},"720P":{"width":1280,"factor":2.2,"height":720},"1080P":{"width":1920,"factor":4.6,"height":1080}}},"inventories":0,"created_at":1779277968000,"updated_at":1779278393000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai-testing.su.bcebos.com/v1/uploads/img/id-m28jg1a44-20260520-163830.jpg","description_html":"ViduQ2-Pro图生视频是全球首创「万物可参考」视频模型。支持特效、表情、纹理、动作、人物、场景等六大维度参考，实现编辑全面进化。通过可控式增、删、改，达成精细化视频编辑，专为漫剧、短剧、影视制作打造的生产级创作引擎
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1191,"name":"视频生成","slug":"video-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text_to_video.svg","en_name":"Video Generation","color":"","description":"","href":"","created_at":1739973857876,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":28210,"name":"ViduQ2-Pro","namespace":"moark","title":"ViduQ2-Pro"},"total_calls":6,"operation_summary":{"service_id":10746,"unit_tag":1264,"min_price":0.1563,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.1563,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":3},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{"540P":{"name":"540P","width":960,"height":540,"price_factor":1.0,"pixels":518400},"720P":{"name":"720P","width":1280,"height":720,"price_factor":2.2,"pixels":921600},"1080P":{"name":"1080P","width":1920,"height":1080,"price_factor":4.6,"pixels":2073600}},"all_vendors":false,"subscription_service":false},"operations":[{"id":750,"account":0,"account_type":0,"type":"text2video","service_id":10746,"service_type":"serverless","model_id":28210,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文生视频","description":"输入文本描述，生成视频。","path":"v1/async/videos/generations","protocol":"http","price":0.1563,"average_price":0.1563,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"540P","value":"540P"},{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"720P"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":10,"minimum":5,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"body","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机数种子，取值范围为[0, 2147483647]。"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1780034617000,"updated_at":1781686429000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[750],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264},{"id":740,"account":0,"account_type":0,"type":"text2video","service_id":10746,"service_type":"serverless","model_id":28210,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"基于首尾帧","description":"输入首帧图片、尾帧图片与文本描述，生成视频。","path":"v1/async/videos/generations","protocol":"http","price":0.1563,"average_price":0.1563,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":1,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"540P","value":"540P"},{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"720P"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":10,"minimum":5,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"body","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机数种子，取值范围为[0, 2147483647]。"},{"in":"body","expr":"$.first_frame","name":"first_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"image"}},"control":"file","category":"main","encoding":"data-url","required":true,"description":"首帧图片"},{"in":"body","expr":"$.last_frame","name":"last_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"image"}},"control":"file","category":"main","encoding":"data-url","required":true,"description":"尾帧图片"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1779277968000,"updated_at":1781686430000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[740],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264},{"id":739,"account":0,"account_type":0,"type":"text2video","service_id":10746,"service_type":"serverless","model_id":28210,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"图生视频","description":"输入首帧图片与文本描述，生成视频。","path":"v1/async/videos/generations","protocol":"http","price":0.1563,"average_price":0.1563,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":2,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"540P","value":"540P"},{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"540P"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":16,"minimum":5,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"body","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机种子"},{"in":"body","expr":"$.first_frame","name":"first_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"image"}},"control":"file","category":"main","encoding":"data-url","required":true,"description":"首帧图片"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1779277968000,"updated_at":1781686430000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[739],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264}]},{"service":{"id":10747,"ident":"ViduQ3-Pro","type":"serverless","vendor":"","region":"","compute":"","model":28208,"name":"ViduQ3-Pro","remark":"ViduQ3-Pro支持长达16秒的音画同步生成，实现多镜头自由切换，精准把控节奏、情绪与叙事连贯性。参数量领先，画质、人物一致性及情绪表现卓越，达电影级标准。适用于广告（电商、TVC、效果投放）、漫剧、真人剧及游戏等专业生产场景。","price":0.0000,"props":{"name":"ViduQ3-Pro","service_type":"serverless","resolution_price_factors":{"540P":{"width":960,"factor":3125.0,"height":540},"720P":{"width":1280,"factor":7813.0,"height":720},"1080P":{"width":1920,"factor":9375.0,"height":1080}}},"inventories":0,"created_at":1779278138000,"updated_at":1779278138000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-6rjp6l3d9-20260520-163007.png","description_html":"ViduQ3-Pro支持长达16秒的音画同步生成，实现多镜头自由切换，精准把控节奏、情绪与叙事连贯性。参数量领先，画质、人物一致性及情绪表现卓越，达电影级标准。适用于广告（电商、TVC、效果投放）、漫剧、真人剧及游戏等专业生产场景。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1191,"name":"视频生成","slug":"video-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text_to_video.svg","en_name":"Video Generation","color":"","description":"","href":"","created_at":1739973857876,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":28208,"name":"ViduQ3-Pro","namespace":"moark","title":"ViduQ3-Pro"},"total_calls":5,"operation_summary":{"service_id":10747,"unit_tag":1264,"min_price":0.0001,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0001,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":3},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{"540P":{"name":"540P","width":960,"height":540,"price_factor":3125.0,"pixels":518400},"720P":{"name":"720P","width":1280,"height":720,"price_factor":7813.0,"pixels":921600},"1080P":{"name":"1080P","width":1920,"height":1080,"price_factor":9375.0,"pixels":2073600}},"all_vendors":false,"subscription_service":false},"operations":[{"id":742,"account":0,"account_type":0,"type":"text2video","service_id":10747,"service_type":"serverless","model_id":28208,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文生视频","description":"Vidu3P 文生视频","path":"v1/async/videos/generations","protocol":"http","price":0.0001,"average_price":0.0001,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":10,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"540P","value":"540P"},{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"720P"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":16,"minimum":5,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"body","expr":"$.audio","name":"audio","type":"boolean","control":"boolean","required":false,"description":"是否生成有声视频。开启后模型将根据视频内容自动生成匹配的背景音乐或音效。","default_value":false},{"in":"body","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机数种子，取值范围为[0, 2147483647]。"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1779278500000,"updated_at":1780034774000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[742],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264},{"id":743,"account":0,"account_type":0,"type":"text2video","service_id":10747,"service_type":"serverless","model_id":28208,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"基于首尾帧","description":"输入首帧图片、尾帧图片与文本描述，生成视频。","path":"v1/async/videos/generations","protocol":"http","price":0.0001,"average_price":0.0001,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":20,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"540P","value":"540P"},{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"720P"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":16,"minimum":5,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"body","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机数种子，取值范围为[0, 2147483647]。"},{"in":"body","expr":"$.first_frame","name":"first_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"image"}},"control":"file","category":"main","encoding":"data-url","required":true,"description":"首帧图片"},{"in":"body","expr":"$.last_frame","name":"last_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"image"}},"control":"file","category":"main","encoding":"data-url","required":true,"description":"尾帧图片"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1779278500000,"updated_at":1780035412000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[743],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264},{"id":741,"account":0,"account_type":0,"type":"text2video","service_id":10747,"service_type":"serverless","model_id":28208,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"图生视频","description":"输入首帧图片与文本描述，生成视频。","path":"v1/async/videos/generations","protocol":"http","price":0.0001,"average_price":0.0001,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":30,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"540P","value":"540P"},{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"540P"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"number","control":"slider","maximum":16,"minimum":5,"required":false,"description":"视频时长（秒）","default_value":5},{"in":"body","expr":"$.seed","name":"seed","type":"number","control":"text","required":false,"description":"随机种子"},{"in":"body","expr":"$.first_frame","name":"first_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"image"}},"control":"file","category":"main","encoding":"data-url","required":true,"description":"首帧图片"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1779278500000,"updated_at":1779348340000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[741],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264}]},{"service":{"id":10901,"ident":"Qwen3.7-Max","type":"serverless","vendor":"","region":"","compute":"","model":28228,"name":"Qwen3.7-Max","remark":"Qwen3.7是面向智能体时代的新一代旗舰模型，核心优势在于智能体能力的广度与深度：在编程、办公与生产力、长周期自主执行方面均能出色胜任各项任务。","price":0.0000,"props":{"name":"Qwen3.7-Max","service_type":"serverless"},"inventories":0,"created_at":1779448877000,"updated_at":1787316080000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":17,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1282,"name":"1M","slug":"context_size_1M","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-ij94b04s1-context_size (1).png","en_name":"1M","color":"blue","description":"上下文大小","href":"","created_at":1781233150090,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28228,"name":"Qwen3.7-Max","namespace":"moark","title":"Qwen3.7-Max"},"total_calls":4,"operation_summary":{"service_id":10901,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":12.00000000,"min_output_million_tokens_price":36.00000000,"max_price":0.0000,"max_input_million_tokens_price":12.00000000,"max_output_million_tokens_price":36.00000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":746,"account":0,"account_type":0,"type":"text2text","service_id":10901,"service_type":"serverless","model_id":28228,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":12.00000000,"output_million_tokens_price":36.00000000,"input_million_tokens_price_average":12.00000000,"output_million_tokens_price_average":36.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":2.4,"cache_write_million_tokens_price":12},"version":1,"specified":{"cache_read_million_tokens_price":2.4,"cache_write_million_tokens_price":12}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1779448993000,"updated_at":1787316098000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[746],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":12107,"ident":"Hy-MT2-30B-A3B","type":"serverless","vendor":"","region":"","compute":"","model":28217,"name":"Hy-MT2-30B-A3B 翻译大模型","remark":"Hy-MT2-30B-A3B 模型是一款高性能多语言翻译大模型，具备优秀的跨语言理解与生成能力，适用于实时翻译、本地化与全球化AI应用场景。","price":0.0000,"props":{"name":"Hy-MT2-30B-A3B","service_type":"serverless"},"inventories":0,"created_at":1779862155000,"updated_at":1779862155000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/news_cover_1-avatar-hunyuan.png","description_html":"\uD83D\uDCCC 模型详细介绍
Hy-MT2-30B-A3B 是一款面向多语言翻译与跨语言理解场景优化的大语言模型，采用高效混合专家架构（MoE），在保证翻译质量的同时兼顾推理效率与部署成本。
模型支持多语种文本翻译、语言理解与上下文语义转换，能够在复杂语境下保持较高的语义一致性与表达自然度，适用于国际化产品、跨境业务及多语言AI服务场景。
除了基础机器翻译能力外，模型还具备较强的上下文理解与内容生成能力，可用于多语言对话、内容本地化、跨语言知识处理等任务。通过标准API接口即可快速集成到翻译平台、AI助手、企业系统及全球化应用中。
\uD83D\uDE80 核心能力
\uD83C\uDF0D 多语言文本翻译
\uD83D\uDCAC 跨语言对话理解
✍ 内容本地化与改写
\uD83E\uDDE0 上下文语义理解
\uD83D\uDCC4 长文本翻译与总结
\uD83D\uDCCA 结构化多语言输出
\uD83C\uDFAF 应用场景
\uD83C\uDF10 国际化网站与应用翻译
\uD83E\uDD16 多语言AI客服与助手
\uD83D\uDCDA 文档与内容本地化
\uD83D\uDECD 跨境电商商品翻译
\uD83C\uDFAE 游戏与软件多语言适配
\uD83E\uDDFE 企业全球化知识库系统
⭐ 模型特色
⚡ 混合专家架构，兼顾性能与成本
\uD83C\uDF0D 优秀的多语言支持能力
\uD83E\uDDE0 强上下文语义一致性
\uD83D\uDE80 适合高并发API服务场景
\uD83D\uDD27 支持OpenAI兼容接口快速接入
\uD83D\uDCCC 模型详细介绍
Hy-MT2-7B 是一款面向多语言翻译与国际化场景优化的轻量级大语言模型，在保证翻译质量的同时兼顾推理速度与部署成本，适合在线API服务与大规模调用场景。
模型支持多语种文本翻译、跨语言语义理解以及内容本地化处理，能够在多种语言之间实现自然流畅的转换，适用于实时翻译、跨境业务、国际化产品以及多语言AI应用。
相比更大规模模型，该模型具备更低资源消耗与更快响应速度，在高并发、低延迟场景下表现优秀，可快速接入各类翻译系统、AI助手与企业服务平台。
\uD83D\uDE80 核心能力
\uD83C\uDF0D 多语言文本翻译
\uD83D\uDCAC 跨语言语义理解
✍ 内容本地化与改写
\uD83D\uDCC4 长文本翻译处理
\uD83D\uDCCA 结构化多语言输出
⚡ 实时翻译响应能力
\uD83C\uDFAF 应用场景
\uD83C\uDF10 国际化网站与应用翻译
\uD83E\uDD16 多语言AI客服与助手
\uD83D\uDECD 跨境电商商品翻译
\uD83D\uDCDA 文档与内容本地化
\uD83C\uDFAE 游戏与软件多语言适配
\uD83D\uDCF1 实时聊天翻译与字幕生成
⭐ 模型特色
⚡ 轻量化架构，响应速度快
\uD83C\uDF0D 多语言支持能力优秀
\uD83D\uDCB0 更低部署与调用成本
\uD83D\uDE80 适合高并发在线API服务
\uD83D\uDD27 支持OpenAI兼容接口快速集成
MiniMax M3 在编码与智能体评测中达到行业顶尖水平，具备自主任务拆解、工具调用与多步推理能力，写出的代码目标是直接可交付，而不是「能跑但需要人改」。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":16,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1163,"name":"多模态","slug":"multi-modal","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287910629,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1282,"name":"1M","slug":"context_size_1M","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-ij94b04s1-context_size (1).png","en_name":"1M","color":"blue","description":"上下文大小","href":"","created_at":1781233150090,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28438,"name":"MiniMax-M3","namespace":"hf-models","mirror_namespace":"MiniMaxAI","mirror_url":"https://huggingface.co/MiniMaxAI/MiniMax-M3"},"total_calls":619,"operation_summary":{"service_id":12943,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":4.20000000,"min_output_million_tokens_price":16.80000000,"max_price":0.0000,"max_input_million_tokens_price":4.20000000,"max_output_million_tokens_price":16.80000000,"free_operation_count":2,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":767,"account":0,"account_type":0,"type":"text2text","service_id":12943,"service_type":"serverless","model_id":28438,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":4.20000000,"output_million_tokens_price":16.80000000,"input_million_tokens_price_average":4.20000000,"output_million_tokens_price_average":16.80000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":512},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":1},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":50,"minimum":0,"required":false,"description":"取值范围为 [0,50]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":1},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":0},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"文件类型：.jpeg, .jpg, .png, .webp，最大体积：5MB"},{"in":"body","name":"messages","type":"string","control":"string","samples":[{"label":"宇宙飞船","value":[{"role":"user","content":[{"text":"这张图的内容是什么？","type":"text"},{"type":"image_url","image_url":{"url":"https://ai-assets.gitee.com/landing-page/example1.png"}}]}]},{"label":"中文诗","value":[{"role":"user","content":[{"text":"图中文字是什么？","type":"text"},{"type":"image_url","image_url":{"url":"https://ai-assets.gitee.com/landing-page/example2.png"}}]}]},{"label":"英文","value":[{"role":"user","content":[{"text":"图片中内容是什么，介绍它的出处，并翻译成中文","type":"text"},{"type":"image_url","image_url":{"url":"https://ai-assets.gitee.com/landing-page/example3.png"}}]}]}],"category":"main","required":true},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1781097937000,"updated_at":1787648973000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[767],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":927,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":28438,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216},{"id":927,"account":0,"account_type":0,"type":"text2text","service_id":12943,"service_type":"serverless","model_id":28438,"vendor":"ztzn","region":"ztzn-region","compute":"ztzn/ztzn-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":4.20000000,"output_million_tokens_price":16.80000000,"input_million_tokens_price_average":4.20000000,"output_million_tokens_price_average":16.80000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":512},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":1},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":50,"minimum":0,"required":false,"description":"取值范围为 [0,50]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":1},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":0},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"文件类型：.jpeg, .jpg, .png, .webp，最大体积：5MB"},{"in":"body","name":"messages","type":"string","control":"string","samples":[{"label":"宇宙飞船","value":[{"role":"user","content":[{"text":"这张图的内容是什么？","type":"text"},{"type":"image_url","image_url":{"url":"https://ai-assets.gitee.com/landing-page/example1.png"}}]}]},{"label":"中文诗","value":[{"role":"user","content":[{"text":"图中文字是什么？","type":"text"},{"type":"image_url","image_url":{"url":"https://ai-assets.gitee.com/landing-page/example2.png"}}]}]},{"label":"英文","value":[{"role":"user","content":[{"text":"图片中内容是什么，介绍它的出处，并翻译成中文","type":"text"},{"type":"image_url","image_url":{"url":"https://ai-assets.gitee.com/landing-page/example3.png"}}]}]}],"category":"main","required":true},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1787649180000,"updated_at":1787649197000,"vendor_info":{"id":0,"name":"ztzn","label":"正拓智能","status":"available","ratio":0.70},"region_info":{"name":"ztzn-region","label":"正拓智能默认","props":{}},"compute_info":{"id":"ztzn/ztzn-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"正拓智能","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"ztzn-region","vendor_name":"ztzn","type":"SERVERLESS_API","storage":0},"ids":[927],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":767,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":28438,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":13019,"ident":"MiMo-V2.5-Pro","type":"serverless","vendor":"","region":"","compute":"","model":28212,"name":"MiMo-V2.5-Pro","remark":"MiMo-V2.5-Pro 是一个开源的专家混合（MoE）语言模型，具有 1.02T 的总参数量和 42B 的活跃参数量。它采用了 MiMo-V2-Flash 中引入的混合注意力架构和 3 层多标记预测（MTP），支持高达 1M 标记的上下文长度。","price":0.0000,"props":{"name":"MiMo-V2.5-Pro","service_type":"serverless"},"inventories":0,"created_at":1780399311000,"updated_at":1787316131000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-ywjwp6tph-5w_4aLfF-7MAyaIPOV498.webp","description_html":"MiMo-V2.5-Pro 是一个开源的专家混合（MoE）语言模型，具有 1.02T 的总参数量和 42B 的活跃参数量。它采用了 MiMo-V2-Flash 中引入的混合注意力架构和 3 层多标记预测（MTP），支持高达 1M 标记的上下文长度。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":15,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28212,"name":"MiMo-V2.5-Pro","namespace":"hf-models","mirror_namespace":"XiaomiMiMo","mirror_url":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro"},"total_calls":2,"operation_summary":{"service_id":13019,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":7.00000000,"min_output_million_tokens_price":21.00000000,"max_price":0.0000,"max_input_million_tokens_price":7.00000000,"max_output_million_tokens_price":21.00000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":762,"account":0,"account_type":0,"type":"text2text","service_id":13019,"service_type":"serverless","model_id":28212,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":7.00000000,"output_million_tokens_price":21.00000000,"input_million_tokens_price_average":7.00000000,"output_million_tokens_price_average":21.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":1.4,"cache_write_million_tokens_price":7},"version":1,"specified":{"cache_read_million_tokens_price":1.4,"cache_write_million_tokens_price":7}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1780399556000,"updated_at":1787316154000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[762],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":13177,"ident":"TeleTTS-Mandarin","type":"serverless","vendor":"teleai","region":"teleai-region","compute":"teleai/teleai-region/serverless_api/gpu/large/default","model":0,"name":"TeleTTS-Mandarin","remark":"可实时将文本转为超自然高品质的语音输出，集情感丰富、自然流畅、多样音色等优势于一身，实现超拟人语音合成，支持普通话版,支持29个音色。","price":0.0000,"props":{"service_type":"serverless"},"inventories":0,"created_at":1780555616000,"updated_at":1781264297000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-b5npuanpa-teleai.jpeg","icon":"","description_html":"可实时将文本转为超自然高品质的语音输出，集情感丰富、自然流畅、多样音色等优势于一身，实现超拟人语音合成，支持普通话版，支持18个音色。
适用于内容播放、语音导航等无需即时生成语音的场景。
基于端到端语音识别能力，可将批量上传的长段音频文件2小时转换成文本数据，为信息快速处理和电子存档提供基础。转写速度可达到1小时音频20秒转写完成。支持中+英+60种方言混合识别。
适用于数字人客服、语音搜索、电话客服、语音导航、智能家居语音指令、手机语音输入等短语音识别场景。
\uD83D\uDCCC 模型详细介绍
MOSS-Audio-8B-Thinking 是一款面向语音理解、多人对话识别与音频推理场景打造的多模态音频大模型。模型不仅支持基础的语音转文本能力，还能够理解音频中的语义信息、上下文关系以及多人对话结构，适用于会议、访谈、客服通话、语音问答等复杂音频场景。
相比传统语音识别模型仅输出文本内容，MOSS-Audio-8B-Thinking 进一步增强了对多人语音场景的理解能力，能够在多角色对话中识别不同说话人的发言内容，理解对话上下文，并基于音频信息进行总结、分析和问答。
该模型可帮助开发者快速构建具备多人对话转写、语音内容理解、音频摘要、会议分析和语音问答能力的 AI 应用，广泛适用于智能客服、会议记录、教育培训、企业知识管理、呼叫中心质检等场景。
\uD83D\uDE80 核心能力
\uD83C\uDF99 多人语音识别与对话转写
支持对多人语音内容进行识别，适用于会议、访谈、客服通话等多角色对话场景。
\uD83D\uDC65 说话人区分与上下文理解
能够理解不同说话人之间的对话关系，辅助生成更清晰的对话记录与内容分析结果。
\uD83E\uDDE0 基于音频的推理与问答
支持根据音频内容进行理解、分析和问答，能够从语音中提取关键信息并生成反馈。
\uD83D\uDCDD 语音转文本与内容提取
可将音频内容转换为文本，并提取重点信息、关键观点、任务事项等内容。
\uD83D\uDCC4 音频摘要与关键信息提炼
支持对会议、课程、播客、通话等长音频进行总结，生成结构化摘要。
\uD83D\uDCAC 多轮语音交互支持
适用于语音助手、智能客服等需要连续对话理解的应用场景。
\uD83D\uDD0D 音频内容检索与分析
可用于企业知识库、录音归档、客服质检等场景中的语音内容检索与分析。
\uD83C\uDFAF 应用场景
\uD83E\uDD16 智能语音助手与 AI 客服
理解用户语音输入，支持自然语音问答和多轮交互。
\uD83D\uDCDE 呼叫中心质检与通话分析
识别客服与用户的对话内容，辅助进行服务质量评估、问题归因和重点信息提取。
\uD83D\uDCCB 多人会议记录与自动总结
适用于多人会议、访谈、讨论等场景，可生成对话记录、会议纪要和待办事项。
\uD83C\uDF93 教育培训与语音问答系统
支持课程音频理解、课堂互动分析、知识点提取和语音问答。
\uD83C\uDFA7 播客、课程与音频内容分析
可对长音频内容进行转写、摘要、重点提炼和内容检索。
\uD83C\uDFE2 企业知识库与语音检索系统
将会议、培训、客服录音等音频资料转化为可检索、可分析的知识内容。
可实时将文本转为超自然高品质的语音输出，集情感丰富、自然流畅、多样音色等优势于一身，实现超拟人语音合成，支持输入输出均为流式，合成支持自动断句（逐字输入），支持29个音色。
适用于智能客服场景。
可实时将文本转为超自然高品质的语音输出，集情感丰富、自然流畅、多样音色等优势于一身，实现超拟人语音合成，支持中+英+60种方言混合识别。
适用于语音质检、政务服务、会议记录等场景。
GLM-5.2是智谱AI面向长程任务（Long Horizon Task）设计的新一代开源旗舰模型，支持1M无损超长上下文。具备卓越的编程与工程能力，可自主完成任务拆解、架构设计、前后端开发、联调测试到多端部署的完整链路，适用于复杂工程任务、长程交互、代码生成、企业应用等场景。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":14,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1269,"name":"128K","slug":"context_size_output_128k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-9y385u6wl-id-7pfkojri0-arrow-down.png","en_name":"128K","color":"blue","description":"最大生成长度","href":"","created_at":1768877784722,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1282,"name":"1M","slug":"context_size_1M","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-ij94b04s1-context_size (1).png","en_name":"1M","color":"blue","description":"上下文大小","href":"","created_at":1781233150090,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28360,"name":"GLM-5.2","namespace":"hf-models","mirror_namespace":"zai-org","mirror_url":"https://huggingface.co/zai-org/GLM-5.2"},"total_calls":9814,"operation_summary":{"service_id":14367,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":8.00000000,"min_output_million_tokens_price":28.00000000,"max_price":0.0000,"max_input_million_tokens_price":8.00000000,"max_output_million_tokens_price":28.00000000,"free_operation_count":2,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":773,"account":0,"account_type":0,"type":"text2text","service_id":14367,"service_type":"serverless","model_id":28360,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":8.00000000,"output_million_tokens_price":28.00000000,"input_million_tokens_price_average":8.00000000,"output_million_tokens_price_average":28.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":2,"cache_write_million_tokens_price":8},"version":1,"specified":{"cache_read_million_tokens_price":2,"cache_write_million_tokens_price":8}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1781664522000,"updated_at":1787554745000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[773],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":788,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":28360,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216},{"id":788,"account":0,"account_type":0,"type":"text2text","service_id":14367,"service_type":"serverless","model_id":28360,"vendor":"ztzn","region":"ztzn-region","compute":"ztzn/ztzn-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":8.00000000,"output_million_tokens_price":28.00000000,"input_million_tokens_price_average":8.00000000,"output_million_tokens_price_average":28.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":2,"cache_write_million_tokens_price":8},"version":1,"specified":{"cache_read_million_tokens_price":2,"cache_write_million_tokens_price":8}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1782992379000,"updated_at":1787556290000,"vendor_info":{"id":0,"name":"ztzn","label":"正拓智能","status":"available","ratio":0.70},"region_info":{"name":"ztzn-region","label":"正拓智能默认","props":{}},"compute_info":{"id":"ztzn/ztzn-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"正拓智能","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"ztzn-region","vendor_name":"ztzn","type":"SERVERLESS_API","storage":0},"ids":[788],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":773,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":28360,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":14452,"ident":"Kimi-K2.7-Code","type":"serverless","vendor":"","region":"","compute":"","model":28363,"name":"Kimi-K2.7-Code","remark":"Kimi-K2.7-Code是一个基于 Kimi K2.6 构建的专注于编码的代理模型。通过对现实世界长期编码任务的显著改进，它增强了复杂软件工程工作流程中的端到端任务完成能力，同时提升了令牌效率，相较于 Kimi K2.6，思考令牌使用率降低了约 30%。","price":0.0000,"props":{"name":"Kimi-K2.7-Code","service_type":"serverless"},"inventories":0,"created_at":1781753433000,"updated_at":1787315987000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-ydbuajvxn-id-2pzepojrt-id-g6laj4iw1-下载.png","description_html":"Kimi-K2.7-Code是一个基于 Kimi K2.6 构建的专注于编码的代理模型。通过对现实世界长期编码任务的显著改进，它增强了复杂软件工程工作流程中的端到端任务完成能力，同时提升了令牌效率，相较于 Kimi K2.6，思考令牌使用率降低了约 30%。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":13,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1228,"name":"256K","slug":"context_size_256k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-9zvdx4rar-id-zd1snxypr-context_size.png","en_name":"256K","color":"blue","description":"上下文大小","href":"","created_at":1753329137839,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28363,"name":"Kimi-K2.7-Code","namespace":"hf-models","mirror_namespace":"moonshotai","mirror_url":"https://huggingface.co/moonshotai/Kimi-K2.7-Code"},"total_calls":45,"operation_summary":{"service_id":14452,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":6.50000000,"min_output_million_tokens_price":27.00000000,"max_price":0.0000,"max_input_million_tokens_price":6.50000000,"max_output_million_tokens_price":27.00000000,"free_operation_count":2,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":774,"account":0,"account_type":0,"type":"text2text","service_id":14452,"service_type":"serverless","model_id":28363,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":6.50000000,"output_million_tokens_price":27.00000000,"input_million_tokens_price_average":6.50000000,"output_million_tokens_price_average":27.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":1.3,"cache_write_million_tokens_price":6.5},"version":1,"specified":{"cache_read_million_tokens_price":1.3,"cache_write_million_tokens_price":6.5}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","type":"number","control":"number","visible":false,"required":false,"description":"temperature","will_be_null":true},{"in":"body","name":"top_p","type":"number","control":"number","visible":false,"required":false,"description":"top_p","will_be_null":true},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","will_be_null":true},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"最大体积：5MB"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1781753582000,"updated_at":1787315977000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[774],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":952,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":28363,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216},{"id":952,"account":0,"account_type":0,"type":"text2text","service_id":14452,"service_type":"serverless","model_id":28363,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":6.50000000,"output_million_tokens_price":27.00000000,"input_million_tokens_price_average":6.50000000,"output_million_tokens_price_average":27.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":1.3,"cache_write_million_tokens_price":6.5},"version":1,"specified":{"cache_read_million_tokens_price":1.3,"cache_write_million_tokens_price":6.5}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","type":"number","control":"number","visible":false,"required":false,"description":"temperature","will_be_null":true},{"in":"body","name":"top_p","type":"number","control":"number","visible":false,"required":false,"description":"top_p","will_be_null":true},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","will_be_null":true},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"最大体积：5MB"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1788417928000,"updated_at":1788417944000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[952],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":774,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":28363,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":14713,"ident":"happyhorse-1.1","type":"serverless","vendor":"","region":"","compute":"","model":28371,"name":"HappyHorse-1.1","remark":"HappyHorse-1.1 模型是一款高质量AI视频生成模型，支持文生视频、图生视频及音视频同步生成，适用于营销创作、短视频制作与数字内容生产场景。","price":0.0000,"props":{"name":"HappyHorse-1.1","service_type":"serverless","resolution_price_factors":{"720P":{"width":1280,"factor":54,"height":720},"1080P":{"width":1920,"factor":72,"height":1080}}},"inventories":0,"created_at":1782116413000,"updated_at":1782122605000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai-testing.su.bcebos.com/v1/uploads/img/id-qgwi3o2fi-id-y0hgtll76-O1CN01K9R3vH1WLAdM94elo_!!6000000002771-2-tps-84-84.png","description_html":"\uD83D\uDCCC 模型详细介绍
HappyHorse-1.1 是新一代多模态视频生成模型，具备从文本或图片直接生成高质量视频内容的能力。模型融合视觉理解、运动生成、语音同步与场景构建能力，能够生成具备自然动作、真实光影效果和流畅镜头衔接的视频作品。
相比传统视频生成模型，HappyHorse 系列更加注重角色一致性、镜头连贯性以及音视频同步效果，可支持多镜头叙事、人物对白、场景切换及多语言口型同步等复杂创作需求。模型能够在一次生成过程中完成视频与音频的协同生成，显著降低视频制作门槛，提高内容生产效率。
对于企业和开发者而言，该模型适合通过API快速接入视频生成能力，可广泛应用于短视频创作、广告营销、电商展示、数字人内容生产及AI影视创作等场景。
\uD83D\uDE80 核心能力
\uD83C\uDFAC 文本生成视频（Text-to-Video）
\uD83D\uDDBC 图片生成视频（Image-to-Video）
\uD83C\uDF99 音视频同步生成
\uD83D\uDC44 多语言口型同步
\uD83C\uDFAD 多镜头故事生成
\uD83C\uDFA8 电影级画面与光影效果
\uD83D\uDEB6 自然人物动作与场景运动生成
\uD83C\uDF0D 多语言内容创作支持
\uD83C\uDFAF 应用场景
\uD83C\uDFA5 AI短视频创作
\uD83D\uDCE2 品牌营销与广告视频生成
\uD83D\uDECD 电商商品展示视频
\uD83E\uDD16 数字人视频制作
\uD83D\uDCDA 教育培训内容生成
\uD83C\uDFAE 游戏宣传视频制作
\uD83C\uDFAC AI影视与剧情短片创作
\uD83C\uDF0D 多语言国际化视频内容生产
⭐ 模型特色
\uD83C\uDF99 原生音视频协同生成能力
\uD83C\uDFAC 支持多镜头连续叙事
\uD83D\uDC44 高精度多语言口型同步
⚡ 生成速度快，适合API在线调用
\uD83C\uDFA8 高质量人物与场景表现
\uD83D\uDE80 适用于商业级视频内容生产与规模化部署
Qwen3.7系列中高性价比Plus模型，在强大文本能力的基础上全面升级了视觉-语言能力，同时保持了在编码、工具使用和生产力工作流方面的完整智能体能力。其核心特色为多模态交互混合智能体能力，能够感知真实世界场景、读取屏幕并操作 GUI、基于视觉参考生成代码、端到端导航移动应用。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":12,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1163,"name":"多模态","slug":"multi-modal","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287910629,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1282,"name":"1M","slug":"context_size_1M","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-ij94b04s1-context_size (1).png","en_name":"1M","color":"blue","description":"上下文大小","href":"","created_at":1781233150090,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1287,"name":"梯度计费","slug":"tiered_pricing","icon":"","en_name":"Tiered Pricing","color":"","description":"","href":"","created_at":1787913527124,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28389,"name":"Qwen3.7-Plus","namespace":"moark","title":"Qwen3.7-Plus"},"total_calls":198,"operation_summary":{"service_id":15015,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":2.00000000,"min_output_million_tokens_price":8.00000000,"max_price":0.0000,"max_input_million_tokens_price":2.00000000,"max_output_million_tokens_price":8.00000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":782,"account":0,"account_type":0,"type":"text2text","service_id":15015,"service_type":"serverless","model_id":28389,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":2.00000000,"output_million_tokens_price":8.00000000,"input_million_tokens_price_average":2.00000000,"output_million_tokens_price_average":8.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"tiered_prompt_cache_tokens","tiers":[{"average":{"input_million_tokens_price":2,"output_million_tokens_price":8,"cache_read_million_tokens_price":0.2,"cache_write_million_tokens_price":2.5},"specified":{"input_million_tokens_price":2,"output_million_tokens_price":8,"cache_read_million_tokens_price":0.2,"cache_write_million_tokens_price":2.5},"max_input_tokens":256000,"min_input_tokens":0},{"average":{"input_million_tokens_price":6,"output_million_tokens_price":24,"cache_read_million_tokens_price":0.6,"cache_write_million_tokens_price":7.5},"specified":{"input_million_tokens_price":6,"output_million_tokens_price":24,"cache_read_million_tokens_price":0.6,"cache_write_million_tokens_price":7.5},"min_input_tokens":256000}],"version":1,"calculation_mode":"whole_order"},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_completion_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":0},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1779767423000,"updated_at":1787913654000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[782],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":15017,"ident":"Qwen3.6-Flash","type":"serverless","vendor":"","region":"","compute":"","model":28390,"name":"Qwen3.6-Flash","remark":"Qwen3.6原生视觉语言系列Flash模型，模型效果相较3.5-Flash显著提升。本模型重点提升agentic coding能力（在多项代码智能体基准上大幅超越前代）、数学推理和代码推理能力；视觉方面在空间智能能力上显著增强，物体定位与目标检测提升尤为突出。","price":0.0000,"props":{"name":"Qwen3.6-Flash","service_type":"serverless"},"inventories":0,"created_at":1779767423000,"updated_at":1787915017000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3.6原生视觉语言系列Flash模型，模型效果相较3.5-Flash显著提升。本模型重点提升agentic coding能力（在多项代码智能体基准上大幅超越前代）、数学推理和代码推理能力；视觉方面在空间智能能力上显著增强，物体定位与目标检测提升尤为突出。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":11,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1163,"name":"多模态","slug":"multi-modal","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287910629,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1282,"name":"1M","slug":"context_size_1M","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-ij94b04s1-context_size (1).png","en_name":"1M","color":"blue","description":"上下文大小","href":"","created_at":1781233150090,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1287,"name":"梯度计费","slug":"tiered_pricing","icon":"","en_name":"Tiered Pricing","color":"","description":"","href":"","created_at":1787913527124,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28390,"name":"Qwen3.6-Flash","namespace":"moark","title":"Qwen3.6-Flash"},"total_calls":5,"operation_summary":{"service_id":15017,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":1.20000000,"min_output_million_tokens_price":7.20000000,"max_price":0.0000,"max_input_million_tokens_price":1.20000000,"max_output_million_tokens_price":7.20000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":781,"account":0,"account_type":0,"type":"text2text","service_id":15017,"service_type":"serverless","model_id":28390,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":1.20000000,"output_million_tokens_price":7.20000000,"input_million_tokens_price_average":1.20000000,"output_million_tokens_price_average":7.20000000,"billing_strategy":"rule_based","billing_policy":{"type":"tiered_prompt_cache_tokens","tiers":[{"average":{"input_million_tokens_price":1.2,"output_million_tokens_price":7.2,"cache_read_million_tokens_price":0.12,"cache_write_million_tokens_price":1.5},"specified":{"input_million_tokens_price":1.2,"output_million_tokens_price":7.2,"cache_read_million_tokens_price":0.12,"cache_write_million_tokens_price":1.5},"max_input_tokens":256000,"min_input_tokens":0},{"average":{"input_million_tokens_price":4.8,"output_million_tokens_price":28.8,"cache_read_million_tokens_price":0.48,"cache_write_million_tokens_price":6},"specified":{"input_million_tokens_price":4.8,"output_million_tokens_price":28.8,"cache_read_million_tokens_price":0.48,"cache_write_million_tokens_price":6},"min_input_tokens":256000}],"version":1,"calculation_mode":"whole_order"},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_completion_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":0},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1779767423000,"updated_at":1787914967000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[781],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":15018,"ident":"Qwen3.5-Flash","type":"serverless","vendor":"","region":"","compute":"","model":28391,"name":"Qwen3.5-Flash","remark":"Qwen3.5原生视觉语言系列Flash模型，基于混合架构设计，融合了线性注意力机制与稀疏混合专家模型，实现了更高的推理效率。模型效果在纯文本与多模态方面相较3系列均实现飞跃式进步；响应速度快，兼具推理速度和性能。\n\n","price":0.0000,"props":{"name":"Qwen3.5-Flash","service_type":"serverless"},"inventories":0,"created_at":1779767423000,"updated_at":1787915070000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3.5原生视觉语言系列Flash模型，基于混合架构设计，融合了线性注意力机制与稀疏混合专家模型，实现了更高的推理效率。模型效果在纯文本与多模态方面相较3系列均实现飞跃式进步；响应速度快，兼具推理速度和性能。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":10,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1163,"name":"多模态","slug":"multi-modal","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287910629,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1282,"name":"1M","slug":"context_size_1M","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-ij94b04s1-context_size (1).png","en_name":"1M","color":"blue","description":"上下文大小","href":"","created_at":1781233150090,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1287,"name":"梯度计费","slug":"tiered_pricing","icon":"","en_name":"Tiered Pricing","color":"","description":"","href":"","created_at":1787913527124,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28391,"name":"Qwen3.5-Flash","namespace":"moark","title":"Qwen3.5-Flash"},"total_calls":2,"operation_summary":{"service_id":15018,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":0.20000000,"min_output_million_tokens_price":2.00000000,"max_price":0.0000,"max_input_million_tokens_price":0.20000000,"max_output_million_tokens_price":2.00000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":780,"account":0,"account_type":0,"type":"text2text","service_id":15018,"service_type":"serverless","model_id":28391,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0.20000000,"output_million_tokens_price":2.00000000,"input_million_tokens_price_average":0.20000000,"output_million_tokens_price_average":2.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"tiered_prompt_cache_tokens","tiers":[{"average":{"input_million_tokens_price":0.2,"output_million_tokens_price":2,"cache_read_million_tokens_price":0.02,"cache_write_million_tokens_price":0.25},"specified":{"input_million_tokens_price":0.2,"output_million_tokens_price":2,"cache_read_million_tokens_price":0.02,"cache_write_million_tokens_price":0.25},"max_input_tokens":128000,"min_input_tokens":0},{"average":{"input_million_tokens_price":0.8,"output_million_tokens_price":8,"cache_read_million_tokens_price":0.08,"cache_write_million_tokens_price":1},"specified":{"input_million_tokens_price":0.8,"output_million_tokens_price":8,"cache_read_million_tokens_price":0.08,"cache_write_million_tokens_price":1},"max_input_tokens":256000,"min_input_tokens":128000},{"average":{"input_million_tokens_price":1.2,"output_million_tokens_price":12,"cache_read_million_tokens_price":0.12,"cache_write_million_tokens_price":1.5},"specified":{"input_million_tokens_price":1.2,"output_million_tokens_price":12,"cache_read_million_tokens_price":0.12,"cache_write_million_tokens_price":1.5},"min_input_tokens":256000}],"version":1,"calculation_mode":"whole_order"},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_completion_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":0},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1779767423000,"updated_at":1787914967000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[780],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":15072,"ident":"qwen-image-2.0","type":"serverless","vendor":"","region":"","compute":"","model":0,"name":"Qwen-Image-2.0","remark":"Qwen-Image-2.0系列加速版模型，实现了图片生成和图片编辑的融合；具备更专业的文字渲染1k token指令支持能力、更细腻的真实质感，细腻刻画写实场景、更强的语义遵循能力。加速版有效实现了模型效果和性能的最佳平衡。","price":0.0000,"props":{"service_type":"serverless"},"inventories":0,"created_at":1782465808000,"updated_at":1782472670000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen-Image-2.0系列加速版模型，实现了图片生成和图片编辑的融合；具备更专业的文字渲染1k token指令支持能力、更细腻的真实质感，细腻刻画写实场景、更强的语义遵循能力。加速版有效实现了模型效果和性能的最佳平衡。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1148,"name":"图像生成与处理","slug":"image_generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-to-image.svg","en_name":"Image Generation","color":"","description":"","href":"","created_at":1734660944141,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"total_calls":7,"operation_summary":{"service_id":15072,"unit_tag":1227,"min_price":0.2000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.2000,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":783,"account":0,"account_type":0,"type":"text2image","service_id":15072,"service_type":"serverless","model_id":0,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"多模态生成","description":"","path":"v1/images/generations","protocol":"http","price":0.2000,"average_price":0.2000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":1,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","label":"提示词","adapter":{"aliyun":{"emit":"message_content","content_type":"text"}},"control":"text","samples":[{"id":"qwen-image-clothing-pose","label":"服装姿势迁移","value":"图1中的女生穿着图2中的黑色裙子按图3的姿势坐下"}],"category":"main","required":true,"description":"描述要生成或编辑的图片内容，可结合参考图说明服装、姿势、主体和场景。"},{"in":"body","expr":"$.images","name":"images","type":"array","label":"参考图","props":{"textField":false,"imageField":{"max_length":3}},"adapter":{"aliyun":{"emit":"message_content","content_type":"image"}},"control":"multimodal","samples":[{"id":"qwen-image-clothing-pose","label":"服装姿势迁移","value":[{"type":"images","value":["https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20250925/thtclx/input1.png","https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20250925/iclsnx/input2.png","https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20250925/gborgw/input3.png"]}]}],"category":"main","encoding":"data-url","required":false,"max_length":3,"min_length":0,"description":"参考图片列表，按提示词中的图1、图2、图3顺序传入。支持上传图片或粘贴图片 URL。"},{"in":"body","expr":"$.n","name":"n","type":"integer","adapter":{"aliyun":{"emit":"parameter"}},"control":"number","maximum":4,"minimum":1,"required":false,"description":"生成图片数量。","default_value":1},{"in":"body","expr":"$.negative_prompt","name":"negative_prompt","type":"string","adapter":{"aliyun":{"emit":"parameter"}},"control":"text","required":false,"description":"不希望出现在图片中的内容。"},{"in":"body","expr":"$.watermark","name":"watermark","type":"boolean","label":"watermark","adapter":{"aliyun":{"emit":"parameter"}},"control":"boolean","required":false,"description":"是否添加水印。","default_value":false},{"in":"body","expr":"$.seed","name":"seed","type":"integer","label":"seed","adapter":{"aliyun":{"emit":"parameter"}},"control":"number","required":false,"description":"用于复现生成结果的随机种子。"},{"in":"body","expr":"$.size","name":"size","type":"string","adapter":{"aliyun":{"emit":"parameter"}},"control":"select","options":[{"label":"1:1 (1024*1024)","value":"1024*1024"},{"label":"4:3 (1152*864)","value":"1152*864"},{"label":"3:4 (864*1152)","value":"864*1152"},{"label":"16:9 (1344*768)","value":"1344*768"},{"label":"9:16 (768*1344)","value":"768*1344"},{"label":"3:2 (1248*832)","value":"1248*832"},{"label":"2:3 (832*1248)","value":"832*1248"},{"label":"3:2 HD (1536*1024)","value":"1536*1024"},{"label":"2:3 HD (1024*1536)","value":"1024*1536"}],"required":false,"description":"设置输出图像的分辨率，格式为宽*高，例如 1024*1536。qwen-image-2.0 系列图像总像素需在 512*512 至 2048*2048 之间。","default_value":"1024*1024"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1782291649000,"updated_at":1782474661000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[783],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1227}]},{"service":{"id":16142,"ident":"Unlimited-OCR","type":"serverless","vendor":"","region":"","compute":"","model":0,"name":"Unlimited-OCR","remark":"Unlimited-OCR 模型是一款面向长文档解析的智能OCR模型，支持高精度文字识别、版面解析与结构化信息提取，适用于文档数字化及智能文档处理场景。","price":0.0000,"props":{"service_type":"serverless"},"inventories":0,"created_at":1783079289000,"updated_at":1783079289000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-681ssg5yg-avatar-baidu.png","description_html":"\uD83D\uDCCC 模型详细介绍
Unlimited-OCR 是一款面向文档解析场景打造的多模态视觉语言模型，支持对图片、扫描件及 PDF 页面进行智能解析，可识别文档中的文本、表格、版面结构、公式等内容，并输出结构化结果，适用于文档数字化、智能办公、知识库构建及 RAG 数据处理等场景。
模型采用 OpenAI 兼容的文本对话接口（Chat Completions） 进行调用，无需使用专门的 OCR 接口。开发者只需按照标准 messages 格式，将文本指令（Prompt）与待解析图片同时传入模型，即可完成文档解析、内容提取及信息理解等任务，便于快速集成到各类 AI 应用中。
调用时，需要在 messages 中同时包含文本内容（type: \"text\"）和图片内容（type: \"image_url\"），其中 文本 Prompt 建议以 <image> 开头，用于引导模型结合图片内容进行分析。例如：
{\n \"messages\": [\n {\n \"role\": \"system\",\n \"content\": \"\"\n },\n {\n \"role\": \"user\",\n \"content\": [\n {\n \"type\": \"text\",\n \"text\": \"<image>document parsing.\"\n },\n {\n \"type\": \"image_url\",\n \"image_url\": {\n \"url\": \"https://gitee.com/realhugh/imgs/raw/master/show1.jpg\"\n }\n }\n ]\n }\n ]\n}模型支持解析合同、发票、报告、论文、表单、截图等多种文档类型，可完成文字识别、版面分析、表格提取、关键信息抽取及内容总结等任务，为开发者提供高精度、易集成的智能文档解析能力。
\uD83D\uDE80 核心能力
\uD83D\uDCC4 长文档 OCR 识别
\uD83E\uDDFE PDF 文档智能解析
\uD83D\uDCCA 表格与复杂版面识别
\uD83D\uDCDD 文本与版面结构化输出
\uD83D\uDD0D 多页文档连续解析
\uD83D\uDCD0 公式、图片及混合内容识别
\uD83E\uDD16 基于自然语言指令的文档理解与信息提取
\uD83C\uDFAF 应用场景
\uD83D\uDCDA 文档数字化与电子档案建设
\uD83C\uDFE2 企业知识库与 RAG 数据构建
\uD83D\uDCC4 合同、报告、论文智能解析
\uD83D\uDCB3 发票、票据及表单识别
\uD83D\uDCD1 PDF 转 Markdown、HTML 或结构化文本
\uD83E\uDD16 智能办公与文档自动化处理
\uD83D\uDD0D 文档检索、内容摘要及关键信息提取
⭐ 模型特色
\uD83D\uDE80 基于标准 Chat Completions 接口，无需专用 OCR API，集成更加简单
\uD83D\uDDBC 支持文本 + 图片多模态输入，符合 OpenAI 兼容调用方式
\uD83D\uDCC4 支持超长文档、复杂版面、表格及公式等内容解析
\uD83E\uDDE0 支持基于 Prompt 的文档理解、内容分析与信息抽取
⚡ 高精度解析与高性能推理，适用于企业级文档处理及高并发 API 服务场景
Step 3.7 Flash 是一个 198B 参数的稀疏专家混合（MoE）视觉语言模型，它结合了一个 196B 参数的语言主干和一个 1.8B 参数的视觉编码器，用于本地的图像理解。该模型专为高频生产工作负载而设计，每个令牌大约激活 11B 个参数，并能够提供高达每秒 400 个令牌的处理能力。Step 3.7 Flash 支持 256k 的上下文窗口。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":9,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1169,"name":"视觉","slug":"vision-language-label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735288027632,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1228,"name":"256K","slug":"context_size_256k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-9zvdx4rar-id-zd1snxypr-context_size.png","en_name":"256K","color":"blue","description":"上下文大小","href":"","created_at":1753329137839,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28283,"name":"Step-3.7-Flash","namespace":"hf-models","mirror_namespace":"unsloth","mirror_url":"https://huggingface.co/unsloth/Step-3.7-Flash"},"total_calls":5,"operation_summary":{"service_id":17618,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":1.35000000,"min_output_million_tokens_price":8.10000000,"max_price":0.0400,"max_input_million_tokens_price":1.35000000,"max_output_million_tokens_price":8.10000000,"free_operation_count":1,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":842,"account":0,"account_type":0,"type":"text2text","service_id":17618,"service_type":"serverless","model_id":28283,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":1.35000000,"output_million_tokens_price":8.10000000,"input_million_tokens_price_average":1.35000000,"output_million_tokens_price_average":8.10000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":0.27,"cache_write_million_tokens_price":1.35},"version":1,"specified":{"cache_read_million_tokens_price":0.27,"cache_write_million_tokens_price":1.35}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_completion_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":0},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1784186908000,"updated_at":1787315714000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[842],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":850,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":28283,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216},{"id":850,"account":0,"account_type":0,"type":"text2text","service_id":17618,"service_type":"serverless","model_id":28283,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0400,"average_price":0.0400,"input_million_tokens_price":1.35000000,"output_million_tokens_price":8.10000000,"input_million_tokens_price_average":1.35000000,"output_million_tokens_price_average":8.10000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_completion_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":0},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"最大体积：5MB"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1784632916000,"updated_at":1787896608000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[850],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":842,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":28283,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":17787,"ident":"Qwen3.5-Plus","type":"serverless","vendor":"","region":"","compute":"","model":0,"name":"Qwen3.5-Plus","remark":"Qwen3.5 Plus 是通义千问新一代高性能通用模型，融合文本与视觉理解能力，在复杂推理、指令遵循、代码生成、工具调用和多语言任务中表现出色，适合智能助手、知识问答及企业级智能体应用。","price":0.0000,"props":{"service_type":"serverless"},"inventories":0,"created_at":1781661600000,"updated_at":1784279236000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3.5 Plus 是通义千问新一代高性能通用模型，融合文本与视觉理解能力，在复杂推理、指令遵循、代码生成、工具调用和多语言任务中表现出色，适合智能助手、知识问答及企业级智能体应用。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"total_calls":7,"operation_summary":{"service_id":17787,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":0.80000000,"min_output_million_tokens_price":4.80000000,"max_price":0.0000,"max_input_million_tokens_price":0.80000000,"max_output_million_tokens_price":4.80000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":843,"account":0,"account_type":0,"type":"text2text","service_id":17787,"service_type":"serverless","model_id":0,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0.80000000,"output_million_tokens_price":4.80000000,"input_million_tokens_price_average":0.80000000,"output_million_tokens_price_average":4.80000000,"billing_strategy":"rule_based","billing_policy":{"type":"tiered_prompt_cache_tokens","tiers":[{"average":{"input_million_tokens_price":0.8,"output_million_tokens_price":4.8,"cache_read_million_tokens_price":0.08,"cache_write_million_tokens_price":1},"specified":{"input_million_tokens_price":0.8,"output_million_tokens_price":4.8,"cache_read_million_tokens_price":0.08,"cache_write_million_tokens_price":1},"max_input_tokens":128000,"min_input_tokens":0},{"average":{"input_million_tokens_price":2,"output_million_tokens_price":12,"cache_read_million_tokens_price":0.2,"cache_write_million_tokens_price":2.5},"specified":{"input_million_tokens_price":2,"output_million_tokens_price":12,"cache_read_million_tokens_price":0.2,"cache_write_million_tokens_price":2.5},"max_input_tokens":256000,"min_input_tokens":128000},{"average":{"input_million_tokens_price":4,"output_million_tokens_price":24,"cache_read_million_tokens_price":0.4,"cache_write_million_tokens_price":5},"specified":{"input_million_tokens_price":4,"output_million_tokens_price":24,"cache_read_million_tokens_price":0.4,"cache_write_million_tokens_price":5},"min_input_tokens":256000}],"version":1,"calculation_mode":"whole_order"},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_completion_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":0},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1784279738000,"updated_at":1787914361000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[843],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":17788,"ident":"Qwen3-VL-32B-Thinking","type":"serverless","vendor":"","region":"","compute":"","model":0,"name":"Qwen3-VL-32B-Thinking","remark":"Qwen3-VL-32B-Thinking 是一款面向复杂多模态推理的视觉语言模型，能够理解图片、文档、图表及视频内容，并通过思考模式完成数学推理、空间分析、视觉问答和多步骤问题求解。","price":0.0000,"props":{"service_type":"serverless"},"inventories":0,"created_at":1781661600000,"updated_at":1784279273000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3-VL-32B-Thinking 是一款面向复杂多模态推理的视觉语言模型，能够理解图片、文档、图表及视频内容，并通过思考模式完成数学推理、空间分析、视觉问答和多步骤问题求解。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1147,"name":"视觉模型","slug":"vision-language","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-vision-language.svg","en_name":"Vision Language","color":"","description":"","href":"","created_at":1734083179939,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"total_calls":7,"operation_summary":{"service_id":17788,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":2.00000000,"min_output_million_tokens_price":20.00000000,"max_price":0.0000,"max_input_million_tokens_price":2.00000000,"max_output_million_tokens_price":20.00000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":844,"account":0,"account_type":0,"type":"text2text","service_id":17788,"service_type":"serverless","model_id":0,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":2.00000000,"output_million_tokens_price":20.00000000,"input_million_tokens_price_average":2.00000000,"output_million_tokens_price_average":20.00000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","type":"number","control":"number","visible":false,"required":false,"description":"temperature","will_be_null":true},{"in":"body","name":"top_p","type":"number","control":"number","visible":false,"required":false,"description":"top_p","will_be_null":true},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","will_be_null":true},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"最大体积：5MB"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1784279738000,"updated_at":1786965040000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[844],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":17789,"ident":"Qwen3-VL-32B-Instruct","type":"serverless","vendor":"","region":"","compute":"","model":0,"name":"Qwen3-VL-32B-Instruct","remark":"Qwen3-VL-32B-Instruct 是一款高性能多模态指令模型，具备图像理解、文档解析、OCR、图表分析、视觉定位和自然语言生成能力，适用于多模态对话、内容分析及智能视觉助手。","price":0.0000,"props":{"service_type":"serverless"},"inventories":0,"created_at":1781661600000,"updated_at":1784279294000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3-VL-32B-Instruct 是一款高性能多模态指令模型，具备图像理解、文档解析、OCR、图表分析、视觉定位和自然语言生成能力，适用于多模态对话、内容分析及智能视觉助手。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1147,"name":"视觉模型","slug":"vision-language","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-vision-language.svg","en_name":"Vision Language","color":"","description":"","href":"","created_at":1734083179939,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"total_calls":7,"operation_summary":{"service_id":17789,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":2.00000000,"min_output_million_tokens_price":8.00000000,"max_price":0.0000,"max_input_million_tokens_price":2.00000000,"max_output_million_tokens_price":8.00000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":845,"account":0,"account_type":0,"type":"text2text","service_id":17789,"service_type":"serverless","model_id":0,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":2.00000000,"output_million_tokens_price":8.00000000,"input_million_tokens_price_average":2.00000000,"output_million_tokens_price_average":8.00000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","type":"number","control":"number","visible":false,"required":false,"description":"temperature","will_be_null":true},{"in":"body","name":"top_p","type":"number","control":"number","visible":false,"required":false,"description":"top_p","will_be_null":true},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","will_be_null":true},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"最大体积：5MB"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1784279738000,"updated_at":1786965040000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[845],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":17790,"ident":"Qwen3-VL-8B-Thinking","type":"serverless","vendor":"","region":"","compute":"","model":0,"name":"Qwen3-VL-8B-Thinking","remark":"Qwen3-VL-8B-Thinking 是一款兼顾部署效率与深度视觉推理能力的多模态模型，支持图片、文档和图表理解，并可通过思考模式处理数学题、空间关系、视觉逻辑及多步骤问答任务。","price":0.0000,"props":{"service_type":"serverless"},"inventories":0,"created_at":1781661600000,"updated_at":1784279311000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3-VL-8B-Thinking 是一款兼顾部署效率与深度视觉推理能力的多模态模型，支持图片、文档和图表理解，并可通过思考模式处理数学题、空间关系、视觉逻辑及多步骤问答任务。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1147,"name":"视觉模型","slug":"vision-language","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-vision-language.svg","en_name":"Vision Language","color":"","description":"","href":"","created_at":1734083179939,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"total_calls":7,"operation_summary":{"service_id":17790,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":0.50000000,"min_output_million_tokens_price":5.00000000,"max_price":0.0000,"max_input_million_tokens_price":0.50000000,"max_output_million_tokens_price":5.00000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":846,"account":0,"account_type":0,"type":"text2text","service_id":17790,"service_type":"serverless","model_id":0,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0.50000000,"output_million_tokens_price":5.00000000,"input_million_tokens_price_average":0.50000000,"output_million_tokens_price_average":5.00000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","type":"number","control":"number","visible":false,"required":false,"description":"temperature","will_be_null":true},{"in":"body","name":"top_p","type":"number","control":"number","visible":false,"required":false,"description":"top_p","will_be_null":true},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","will_be_null":true},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"最大体积：5MB"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1784279738000,"updated_at":1786965040000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[846],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":17791,"ident":"Qwen3-VL-235B-A22B-Thinking","type":"serverless","vendor":"","region":"","compute":"","model":0,"name":"Qwen3-VL-235B-A22B-Thinking","remark":"Qwen3-VL-235B-A22B-Thinking 是 Qwen3-VL 系列的大规模混合专家思考模型，具备强大的视觉感知、跨模态推理、复杂文档理解、视频分析和智能体能力，适用于高难度多模态推理任务。","price":0.0000,"props":{"service_type":"serverless"},"inventories":0,"created_at":1781661600000,"updated_at":1784279363000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3-VL-235B-A22B-Thinking 是 Qwen3-VL 系列的大规模混合专家思考模型，具备强大的视觉感知、跨模态推理、复杂文档理解、视频分析和智能体能力，适用于高难度多模态推理任务。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1147,"name":"视觉模型","slug":"vision-language","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-vision-language.svg","en_name":"Vision Language","color":"","description":"","href":"","created_at":1734083179939,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"total_calls":7,"operation_summary":{"service_id":17791,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":2.00000000,"min_output_million_tokens_price":20.00000000,"max_price":0.0000,"max_input_million_tokens_price":2.00000000,"max_output_million_tokens_price":20.00000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":847,"account":0,"account_type":0,"type":"text2text","service_id":17791,"service_type":"serverless","model_id":0,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":2.00000000,"output_million_tokens_price":20.00000000,"input_million_tokens_price_average":2.00000000,"output_million_tokens_price_average":20.00000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","type":"number","control":"number","visible":false,"required":false,"description":"temperature","will_be_null":true},{"in":"body","name":"top_p","type":"number","control":"number","visible":false,"required":false,"description":"top_p","will_be_null":true},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","will_be_null":true},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"最大体积：5MB"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1784279738000,"updated_at":1786965040000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[847],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":17792,"ident":"Qwen3-VL-235B-A22B-Instruct","type":"serverless","vendor":"","region":"","compute":"","model":0,"name":"Qwen3-VL-235B-A22B-Instruct","remark":"Qwen3-VL-235B-A22B-Instruct 是 Qwen3-VL 系列的大规模混合专家指令模型，能够高质量理解图片、视频、文档和图表，并完成视觉问答、OCR、信息提取、内容生成及多模态智能体任务。","price":0.0000,"props":{"service_type":"serverless"},"inventories":0,"created_at":1781661600000,"updated_at":1787315649000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3-VL-235B-A22B-Instruct 是 Qwen3-VL 系列的大规模混合专家指令模型，能够高质量理解图片、视频、文档和图表，并完成视觉问答、OCR、信息提取、内容生成及多模态智能体任务。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1147,"name":"视觉模型","slug":"vision-language","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-vision-language.svg","en_name":"Vision Language","color":"","description":"","href":"","created_at":1734083179939,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"total_calls":7,"operation_summary":{"service_id":17792,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":2.00000000,"min_output_million_tokens_price":8.00000000,"max_price":0.0000,"max_input_million_tokens_price":2.00000000,"max_output_million_tokens_price":8.00000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":848,"account":0,"account_type":0,"type":"text2text","service_id":17792,"service_type":"serverless","model_id":0,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":2.00000000,"output_million_tokens_price":8.00000000,"input_million_tokens_price_average":2.00000000,"output_million_tokens_price_average":8.00000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","type":"number","control":"number","visible":false,"required":false,"description":"temperature","will_be_null":true},{"in":"body","name":"top_p","type":"number","control":"number","visible":false,"required":false,"description":"top_p","will_be_null":true},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","will_be_null":true},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"最大体积：5MB"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1784279738000,"updated_at":1786965040000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[848],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":17793,"ident":"Qwen3-Coder-Flash","type":"serverless","vendor":"","region":"","compute":"","model":0,"name":"Qwen3-Coder-Flash","remark":"Qwen3-Coder-Flash 是一款面向高频编程任务优化的代码模型，兼顾代码生成质量与响应速度，支持代码补全、调试、重构、仓库级理解、工具调用和智能体编程，适合在线开发助手及自动化研发工作流。","price":0.0000,"props":{"service_type":"serverless"},"inventories":0,"created_at":1781661600000,"updated_at":1787915112000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3-Coder-Flash 是一款面向高频编程任务优化的代码模型，兼顾代码生成质量与响应速度，支持代码补全、调试、重构、仓库级理解、工具调用和智能体编程，适合在线开发助手及自动化研发工作流。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":8,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1287,"name":"梯度计费","slug":"tiered_pricing","icon":"","en_name":"Tiered Pricing","color":"","description":"","href":"","created_at":1787913527124,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"total_calls":7,"operation_summary":{"service_id":17793,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":1.00000000,"min_output_million_tokens_price":4.00000000,"max_price":0.0000,"max_input_million_tokens_price":1.00000000,"max_output_million_tokens_price":4.00000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":849,"account":0,"account_type":0,"type":"text2text","service_id":17793,"service_type":"serverless","model_id":0,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":1.00000000,"output_million_tokens_price":4.00000000,"input_million_tokens_price_average":1.00000000,"output_million_tokens_price_average":4.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"tiered_prompt_cache_tokens","tiers":[{"average":{"input_million_tokens_price":1,"output_million_tokens_price":4,"cache_read_million_tokens_price":0.2},"specified":{"input_million_tokens_price":1,"output_million_tokens_price":4,"cache_read_million_tokens_price":0.2},"max_input_tokens":32000,"min_input_tokens":0},{"average":{"input_million_tokens_price":1.5,"output_million_tokens_price":6,"cache_read_million_tokens_price":0.3},"specified":{"input_million_tokens_price":1.5,"output_million_tokens_price":6,"cache_read_million_tokens_price":0.3},"max_input_tokens":128000,"min_input_tokens":32000},{"average":{"input_million_tokens_price":2.5,"output_million_tokens_price":10,"cache_read_million_tokens_price":0.5},"specified":{"input_million_tokens_price":2.5,"output_million_tokens_price":10,"cache_read_million_tokens_price":0.5},"max_input_tokens":256000,"min_input_tokens":128000},{"average":{"input_million_tokens_price":5,"output_million_tokens_price":25,"cache_read_million_tokens_price":1},"specified":{"input_million_tokens_price":5,"output_million_tokens_price":25,"cache_read_million_tokens_price":1},"min_input_tokens":256000}],"version":1,"calculation_mode":"whole_order"},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1784279738000,"updated_at":1787914967000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[849],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":18314,"ident":"kimi-k3","type":"serverless","vendor":"","region":"","compute":"","model":28568,"name":"Kimi-K3","remark":"Kimi K3 旗舰多模态推理模型，原生支持视觉理解、深度思考与 100 万 token 上下文。","price":0.0000,"props":{"name":"kimi-k3","service_type":"serverless"},"inventories":0,"created_at":1784700207000,"updated_at":1787315493000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-eup915vlt-裁剪的圆形图片.png","description_html":"Kimi K3 是 Kimi 迄今能力最强的旗舰模型，拥有 2.8 万亿参数，基于 KDA 混合线性注意力机制（Kimi Delta Attention）和注意力残差（Attention Residuals）技术构建，原生支持视觉理解，并拥有 100 万 token 上下文窗口。它是全球首个开源的 3 万亿级别模型，面向长程编程、知识工作和推理等前沿智能场景而设计。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":7,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1169,"name":"视觉","slug":"vision-language-label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735288027632,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1179,"name":"深度思考","slug":"deep_reflection_label","icon":"","en_name":"deep reflection","color":"blue","description":"","href":"","created_at":1737026205631,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1282,"name":"1M","slug":"context_size_1M","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-ij94b04s1-context_size (1).png","en_name":"1M","color":"blue","description":"上下文大小","href":"","created_at":1781233150090,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28568,"name":"kimi-k3","namespace":"moark","title":"kimi-k3"},"total_calls":2827,"operation_summary":{"service_id":18314,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":20.00000000,"min_output_million_tokens_price":100.00000000,"max_price":0.0000,"max_input_million_tokens_price":20.00000000,"max_output_million_tokens_price":100.00000000,"free_operation_count":2,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":852,"account":0,"account_type":0,"type":"text2text","service_id":18314,"service_type":"serverless","model_id":28568,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"Kimi K3 百炼多模态文本对话","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":20.00000000,"output_million_tokens_price":100.00000000,"input_million_tokens_price_average":20.00000000,"output_million_tokens_price_average":100.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":2},"version":1,"specified":{"cache_read_million_tokens_price":2}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","type":"number","control":"number","visible":false,"required":false,"description":"temperature","will_be_null":true},{"in":"body","name":"top_p","type":"number","control":"number","visible":false,"required":false,"description":"top_p","will_be_null":true},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","will_be_null":true},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"最大体积：5MB"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1784700315000,"updated_at":1788232888000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[852],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":922,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":28568,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216},{"id":922,"account":0,"account_type":0,"type":"text2text","service_id":18314,"service_type":"serverless","model_id":28568,"vendor":"ztzn","region":"ztzn-region","compute":"ztzn/ztzn-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":20.00000000,"output_million_tokens_price":100.00000000,"input_million_tokens_price_average":20.00000000,"output_million_tokens_price_average":100.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":2},"version":1,"specified":{"cache_read_million_tokens_price":2}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1787541134000,"updated_at":1787915408000,"vendor_info":{"id":0,"name":"ztzn","label":"正拓智能","status":"available","ratio":0.70},"region_info":{"name":"ztzn-region","label":"正拓智能默认","props":{}},"compute_info":{"id":"ztzn/ztzn-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"正拓智能","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"ztzn-region","vendor_name":"ztzn","type":"SERVERLESS_API","storage":0},"ids":[922],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":852,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":28568,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":18327,"ident":"wan2.7-image","type":"serverless","vendor":"","region":"","compute":"","model":0,"name":"Wan2.7-Image","remark":"万相 2.7 图像生成与编辑模型，支持组图、多图参考、交互式编辑和最高 2K 输出。","price":0.0000,"props":{"service_type":"serverless"},"inventories":0,"created_at":1784704972000,"updated_at":1784794358000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"万相2.7-图像生成与编辑，支持文生图、文生组图、图生组图、图像编辑、多图参考生成、交互式编辑，在文字渲染、主体一致性、复杂指令遵循上都有更强表现
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1148,"name":"图像生成与处理","slug":"image_generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-to-image.svg","en_name":"Image Generation","color":"","description":"","href":"","created_at":1734660944141,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"total_calls":7,"operation_summary":{"service_id":18327,"unit_tag":1227,"min_price":0.2000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.2000,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":855,"account":0,"account_type":0,"type":"text2image","service_id":18327,"service_type":"serverless","model_id":0,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"多模态生成","description":"Wan2.7-Image 百炼图片生成与编辑","path":"v1/images/generations","protocol":"http","price":0.2000,"average_price":0.2000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","billing_policy":{},"sort_order":1,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","label":"提示词","adapter":{"aliyun":{"emit":"message_content","content_type":"text"}},"control":"text","category":"main","required":true,"max_length":5000,"description":"描述要生成或编辑的图片内容。"},{"in":"body","expr":"$.images","name":"images","type":"array","label":"参考图","props":{"textField":false,"imageField":{"max_length":9}},"adapter":{"aliyun":{"emit":"message_content","content_type":"image"}},"control":"multimodal","category":"main","encoding":"data-url","required":false,"max_length":9,"min_length":0,"description":"参考图片列表，最多 9 张；支持上传图片或粘贴图片 URL。"},{"in":"body","expr":"$.n","name":"n","type":"integer","adapter":{"aliyun":{"emit":"parameter"}},"control":"number","maximum":4,"minimum":1,"required":false,"description":"生成图片数量，取值 1 到 4。","default_value":1},{"in":"body","expr":"$.negative_prompt","name":"negative_prompt","type":"string","adapter":{"aliyun":{"emit":"parameter"}},"control":"text","required":false,"description":"不希望出现在图片中的内容。"},{"in":"body","expr":"$.watermark","name":"watermark","type":"boolean","label":"watermark","adapter":{"aliyun":{"emit":"parameter"}},"control":"boolean","visible":false,"required":false,"description":"是否添加水印。","default_value":false},{"in":"body","expr":"$.seed","name":"seed","type":"integer","label":"seed","adapter":{"aliyun":{"emit":"parameter"}},"control":"number","required":false,"description":"用于复现生成结果的随机种子。"},{"in":"body","expr":"$.size","name":"size","type":"string","adapter":{"aliyun":{"emit":"parameter"}},"control":"text","required":false,"description":"输出分辨率档位，支持 1K、2K。","default_value":"2K"},{"in":"body","expr":"$.prompt_extend","name":"prompt_extend","type":"boolean","adapter":{"aliyun":{"emit":"parameter"}},"control":"boolean","required":false,"description":"是否启用智能提示词改写。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1784704972000,"updated_at":1788346678000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[855],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1227},{"id":857,"account":0,"account_type":0,"type":"text2image","service_id":18327,"service_type":"serverless","model_id":0,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"多模态生成（组图模式）","description":"Wan2.7-Image 百炼组图生成；实际返回数量不超过 n","path":"v1/images/generations","protocol":"http","price":0.2000,"average_price":0.2000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","billing_policy":{},"sort_order":1,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","label":"提示词","adapter":{"aliyun":{"emit":"message_content","content_type":"text"}},"control":"text","category":"main","required":true,"max_length":5000,"description":"描述要生成或编辑的图片内容。"},{"in":"body","expr":"$.images","name":"images","type":"array","label":"参考图","props":{"textField":false,"imageField":{"max_length":9}},"adapter":{"aliyun":{"emit":"message_content","content_type":"image"}},"control":"multimodal","category":"main","encoding":"data-url","required":false,"max_length":9,"min_length":0,"description":"参考图片列表，最多 9 张；支持上传图片或粘贴图片 URL。"},{"in":"body","expr":"$.n","name":"n","type":"integer","adapter":{"aliyun":{"emit":"parameter"}},"control":"number","maximum":4,"minimum":1,"required":false,"description":"组图模式最大生成数量，取值 1 到 4，默认 4；实际返回数量由模型决定且不超过 n。","default_value":4},{"in":"body","expr":"$.negative_prompt","name":"negative_prompt","type":"string","adapter":{"aliyun":{"emit":"parameter"}},"control":"text","required":false,"description":"不希望出现在图片中的内容。"},{"in":"body","expr":"$.watermark","name":"watermark","type":"boolean","label":"watermark","adapter":{"aliyun":{"emit":"parameter"}},"control":"boolean","visible":false,"required":false,"description":"是否添加水印。","default_value":false},{"in":"body","expr":"$.seed","name":"seed","type":"integer","label":"seed","adapter":{"aliyun":{"emit":"parameter"}},"control":"number","required":false,"description":"用于复现生成结果的随机种子。"},{"in":"body","expr":"$.size","name":"size","type":"string","adapter":{"aliyun":{"emit":"parameter"}},"control":"text","required":false,"description":"输出分辨率档位，支持 1K、2K。","default_value":"2K"},{"in":"body","expr":"$.prompt_extend","name":"prompt_extend","type":"boolean","adapter":{"aliyun":{"emit":"parameter"}},"control":"boolean","required":false,"description":"是否启用智能提示词改写。","default_value":true},{"in":"body","expr":"$.enable_sequential","name":"enable_sequential","type":"boolean","adapter":{"aliyun":{"emit":"parameter"}},"control":"boolean","visible":false,"category":"main","required":false,"description":"启用组图模式；必须显式传 true。","fixed_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1784709865000,"updated_at":1788346965000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[857],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1227}]},{"service":{"id":18328,"ident":"wan2.7-image-pro","type":"serverless","vendor":"","region":"","compute":"","model":28691,"name":"Wan2.7-Image-Pro","remark":"万相 2.7 图像生成与编辑旗舰版，支持组图、多图参考、交互式编辑和最高 4K 输出。","price":0.0000,"props":{"name":"Wan2.7-Image-Pro","service_type":"serverless"},"inventories":0,"created_at":1784704972000,"updated_at":1787305186000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"万相2.7-图像生成与编辑旗舰版模型，支持文生图、文生组图、图生组图、图像编辑、多图参考生成、交互式编辑，在文字渲染、主体一致性、复杂指令遵循上都有更强表现。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1148,"name":"图像生成与处理","slug":"image_generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-to-image.svg","en_name":"Image Generation","color":"","description":"","href":"","created_at":1734660944141,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":28691,"name":"Wan2.7-Image-Pro","namespace":"moark","title":"Wan2.7-Image-Pro"},"total_calls":3,"operation_summary":{"service_id":18328,"unit_tag":1227,"min_price":0.5000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.5000,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":854,"account":0,"account_type":0,"type":"text2image","service_id":18328,"service_type":"serverless","model_id":28691,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"多模态生成","description":"Wan2.7-Image-Pro 百炼图片生成与编辑","path":"v1/images/generations","protocol":"http","price":0.5000,"average_price":0.5000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","billing_policy":{},"sort_order":1,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","label":"提示词","adapter":{"aliyun":{"emit":"message_content","content_type":"text"}},"control":"text","category":"main","required":true,"max_length":5000,"description":"描述要生成或编辑的图片内容。"},{"in":"body","expr":"$.images","name":"images","type":"array","label":"参考图","props":{"textField":false,"imageField":{"max_length":9}},"adapter":{"aliyun":{"emit":"message_content","content_type":"image"}},"control":"multimodal","category":"main","encoding":"data-url","required":false,"max_length":9,"min_length":0,"description":"参考图片列表，最多 9 张；支持上传图片或粘贴图片 URL。"},{"in":"body","expr":"$.n","name":"n","type":"integer","adapter":{"aliyun":{"emit":"parameter"}},"control":"number","maximum":4,"minimum":1,"required":false,"description":"生成图片数量，取值 1 到 4。","default_value":1},{"in":"body","expr":"$.negative_prompt","name":"negative_prompt","type":"string","adapter":{"aliyun":{"emit":"parameter"}},"control":"text","required":false,"description":"不希望出现在图片中的内容。"},{"in":"body","expr":"$.seed","name":"seed","type":"integer","label":"seed","adapter":{"aliyun":{"emit":"parameter"}},"control":"number","required":false,"description":"用于复现生成结果的随机种子。"},{"in":"body","expr":"$.size","name":"size","type":"string","adapter":{"aliyun":{"emit":"parameter"}},"control":"text","required":false,"description":"输出分辨率档位，支持 1K、2K。","default_value":"2K"},{"in":"body","expr":"$.prompt_extend","name":"prompt_extend","type":"boolean","adapter":{"aliyun":{"emit":"parameter"}},"control":"boolean","required":false,"description":"是否启用智能提示词改写。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1784704972000,"updated_at":1788346980000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[854],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1227},{"id":856,"account":0,"account_type":0,"type":"text2image","service_id":18328,"service_type":"serverless","model_id":28691,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"多模态生成（组图模式）","description":"Wan2.7-Image-Pro 百炼组图生成；实际返回数量不超过 n","path":"v1/images/generations","protocol":"http","price":0.5000,"average_price":0.5000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","billing_policy":{},"sort_order":1,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","label":"提示词","adapter":{"aliyun":{"emit":"message_content","content_type":"text"}},"control":"text","category":"main","required":true,"max_length":5000,"description":"描述要生成或编辑的图片内容。"},{"in":"body","expr":"$.images","name":"images","type":"array","label":"参考图","props":{"textField":false,"imageField":{"max_length":9}},"adapter":{"aliyun":{"emit":"message_content","content_type":"image"}},"control":"multimodal","category":"main","encoding":"data-url","required":false,"max_length":9,"min_length":0,"description":"参考图片列表，最多 9 张；支持上传图片或粘贴图片 URL。"},{"in":"body","expr":"$.n","name":"n","type":"integer","adapter":{"aliyun":{"emit":"parameter"}},"control":"number","maximum":4,"minimum":1,"required":false,"description":"组图模式最大生成数量，取值 1 到 4，默认 4；实际返回数量由模型决定且不超过 n。","default_value":4},{"in":"body","expr":"$.negative_prompt","name":"negative_prompt","type":"string","adapter":{"aliyun":{"emit":"parameter"}},"control":"text","required":false,"description":"不希望出现在图片中的内容。"},{"in":"body","expr":"$.seed","name":"seed","type":"integer","label":"seed","adapter":{"aliyun":{"emit":"parameter"}},"control":"number","required":false,"description":"用于复现生成结果的随机种子。"},{"in":"body","expr":"$.size","name":"size","type":"string","adapter":{"aliyun":{"emit":"parameter"}},"control":"text","required":false,"description":"输出分辨率档位，支持 1K、2K。","default_value":"2K"},{"in":"body","expr":"$.prompt_extend","name":"prompt_extend","type":"boolean","adapter":{"aliyun":{"emit":"parameter"}},"control":"boolean","required":false,"description":"是否启用智能提示词改写。","default_value":true},{"in":"body","expr":"$.enable_sequential","name":"enable_sequential","type":"boolean","adapter":{"aliyun":{"emit":"parameter"}},"control":"boolean","visible":true,"category":"extra","required":true,"description":"启用组图模式；必须显式传 true。","fixed_value":true,"default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1784709865000,"updated_at":1788346997000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[856],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1227}]},{"service":{"id":18329,"ident":"qwen-image-2.0-pro","type":"serverless","vendor":"","region":"","compute":"","model":0,"name":"Qwen-Image-2.0-Pro","remark":"Qwen-Image-2.0 满血版，支持图片生成与编辑、专业文字渲染、多图参考和高分辨率输出。","price":0.0000,"props":{"service_type":"serverless"},"inventories":0,"created_at":1784704972000,"updated_at":1784705022000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen-Image-2.0系列加速版模型，实现了图片生成和图片编辑的融合；具备更专业的文字渲染1k token指令支持能力、更细腻的真实质感，细腻刻画写实场景、更强的语义遵循能力。加速版有效实现了模型效果和性能的最佳平衡。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1148,"name":"图像生成与处理","slug":"image_generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-to-image.svg","en_name":"Image Generation","color":"","description":"","href":"","created_at":1734660944141,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"total_calls":7,"operation_summary":{"service_id":18329,"unit_tag":1227,"min_price":0.5000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.5000,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":853,"account":0,"account_type":0,"type":"text2image","service_id":18329,"service_type":"serverless","model_id":0,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"多模态生成","description":"Qwen-Image-2.0-Pro 百炼图片生成与编辑","path":"v1/images/generations","protocol":"http","price":0.5000,"average_price":0.5000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":1,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","label":"提示词","adapter":{"aliyun":{"emit":"message_content","content_type":"text"}},"control":"text","samples":[{"id":"qwen-image-clothing-pose","label":"服装姿势迁移","value":"图1中的女生穿着图2中的黑色裙子按图3的姿势坐下"}],"category":"main","required":true,"max_length":2000,"description":"描述要生成或编辑的图片内容，可结合参考图说明服装、姿势、主体和场景。"},{"in":"body","expr":"$.images","name":"images","type":"array","label":"参考图","props":{"textField":false,"imageField":{"max_length":3}},"adapter":{"aliyun":{"emit":"message_content","content_type":"image"}},"control":"multimodal","samples":[{"id":"qwen-image-clothing-pose","label":"服装姿势迁移","value":[{"type":"images","value":["https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20250925/thtclx/input1.png","https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20250925/iclsnx/input2.png","https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20250925/gborgw/input3.png"]}]}],"category":"main","encoding":"data-url","required":false,"max_length":3,"min_length":0,"description":"参考图片列表，按提示词中的图1、图2、图3顺序传入。支持上传图片或粘贴图片 URL。"},{"in":"body","expr":"$.n","name":"n","type":"integer","adapter":{"aliyun":{"emit":"parameter"}},"control":"number","maximum":6,"minimum":1,"required":false,"description":"生成图片数量。","default_value":1},{"in":"body","expr":"$.negative_prompt","name":"negative_prompt","type":"string","adapter":{"aliyun":{"emit":"parameter"}},"control":"text","required":false,"max_length":8000,"description":"不希望出现在图片中的内容。"},{"in":"body","expr":"$.watermark","name":"watermark","type":"boolean","label":"watermark","adapter":{"aliyun":{"emit":"parameter"}},"control":"boolean","required":false,"description":"是否添加水印。","default_value":false},{"in":"body","expr":"$.seed","name":"seed","type":"integer","label":"seed","adapter":{"aliyun":{"emit":"parameter"}},"control":"number","required":false,"description":"用于复现生成结果的随机种子。"},{"in":"body","expr":"$.size","name":"size","type":"string","adapter":{"aliyun":{"emit":"parameter"}},"control":"select","options":[{"label":"16:9 (2688*1536)","value":"2688*1536"},{"label":"4:3 (2368*1728)","value":"2368*1728"},{"label":"1:1 (2048*2048)","value":"2048*2048"},{"label":"3:4 (1728*2368)","value":"1728*2368"},{"label":"9:16 (1536*2688)","value":"1536*2688"}],"required":false,"description":"Qwen-Image-2.0-Pro 支持的高分辨率输出尺寸。","default_value":"2048*2048"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1784704972000,"updated_at":1784705568000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[853],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1227}]},{"service":{"id":18480,"ident":"MinerU2.5-Pro","type":"serverless","vendor":"","region":"","compute":"","model":28263,"name":"MinerU2.5-Pro","remark":"MinerU2.5-Pro是一款面向复杂文档解析与 OCR 场景的专业模型，能够从图片和文档中识别文字、理解页面布局，并将非结构化内容转换为便于存储、检索和二次处理的结构化结果。","price":0.0000,"props":{"name":"MinerU2.5-Pro-2604-1.2B","service_type":"serverless"},"inventories":0,"created_at":1784806764000,"updated_at":1784807073000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-spiywsbrp-id-btfubx0rb-avatar-mineru.png","description_html":"模型简介
MinerU2.5-Pro是一款面向复杂文档解析与 OCR 场景的专业模型，能够从图片和文档中识别文字、理解页面布局，并将非结构化内容转换为便于存储、检索和二次处理的结构化结果。
核心能力
高精度识别中文、英文、数字及常见符号
支持标题、段落、列表等文本结构解析
支持表格、公式及复杂版面内容识别
支持多栏排版、图文混排等复杂文档
支持文档阅读顺序和内容层级还原
可将解析结果转换为 Markdown 等结构化格式
上下文长度支持 8K
适用场景
适用于合同、报告、论文、教材、说明书、票据、扫描件及其他复杂文档的自动化解析，可用于知识库构建、文档数字化、内容检索、信息抽取和智能问答等业务场景。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1154,"name":"8K","slug":"context_size_8k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-1e5w7xxiu-context_size.png","en_name":"8K","color":"blue","description":"上下文大小","href":"","created_at":1734693887177,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1160,"name":"多语言","slug":"multi-language-label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287760476,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1175,"name":"文档处理/OCR","slug":"document_processing_label","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-document.svg","en_name":"Document/OCR","color":"","description":"","href":"","created_at":1735902954339,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1178,"name":"文字识别","slug":"text_recognition_label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1736851030007,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28263,"name":"MinerU2.5-Pro-2604-1.2B","namespace":"hf-models","mirror_namespace":"opendatalab","mirror_url":"https://huggingface.co/opendatalab/MinerU2.5-Pro-2604-1.2B"},"total_calls":5,"operation_summary":{"service_id":18480,"unit_tag":1203,"min_price":0.0200,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0200,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":863,"account":0,"account_type":0,"type":"doc2md","service_id":18480,"service_type":"serverless","model_id":28263,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"文档解析","description":"","path":"v1/async/documents/parse","protocol":"http","price":0.0200,"average_price":0.0200,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"form","name":"is_ocr","type":"boolean","control":"boolean","required":false,"description":"是否启用 ocr","default_value":true},{"in":"form","name":"include_image_base64","type":"boolean","control":"boolean","required":false,"description":"启用后，markdown 内嵌 base64 图片, 否则上传到云存储提供临时链接 7 日有效","default_value":true},{"in":"form","name":"formula_enable","type":"boolean","control":"boolean","required":false,"description":"是否启用公式解析","default_value":true},{"in":"form","name":"table_enable","type":"boolean","control":"boolean","required":false,"description":"是否启用表格解析","default_value":true},{"in":"form","name":"end_pages","type":"number","control":"number","required":false,"description":"处理的页数"},{"in":"form","name":"layout_model","type":"string","control":"select","options":[{"label":"layoutlmv3","value":"layoutlmv3"},{"label":"doclayout_yolo","value":"doclayout_yolo"}],"required":false,"description":"布局分析模型","default_value":"doclayout_yolo"},{"in":"form","name":"file","type":"file","accept":".pdf, .png, .jpeg, .jp2, .webp, .gif, .bmp, .jpg","control":"file","category":"main","required":true,"mediaType":"file"},{"in":"form","name":"language","type":"string","control":"text","required":false,"description":"指定文字语言用于提高识别精确度，默认为自动识别。可选语言：ch', 'ch_server', 'ch_lite', 'en', 'korean', 'japan', 'chinese_cht', 'ta', 'te', 'ka', 'th', 'el','latin', 'arabic', 'east_slavic', 'cyrillic', 'devanagari'"},{"in":"form","name":"output_format","type":"string","control":"select","options":[{"label":"md","value":"md"},{"label":"content_list","value":"content_list"},{"label":"middle","value":"middle"}],"required":false,"description":"输出格式","default_value":"md"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1784806842000,"updated_at":1784807017000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[863],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":1203}]},{"service":{"id":19722,"ident":"qwen3.8-max","type":"serverless","vendor":"","region":"","compute":"","model":28583,"name":"Qwen3.8-Max","remark":"2.4万亿参数MoE旗舰，编程与办公能力全面跃升，可自主编程十数天交付完整项目。胜任法律、金融、设计等数百种专业任务，一次对话端到端交付生产级成果。原生视觉理解贯穿规划、执行与验证全流程，支持超长文档与长视频的深度语义解析。长程任务中自主规划与闭环迭代，持续进化。","price":0.0000,"props":{"name":"qwen3.8-max","service_type":"serverless"},"inventories":0,"created_at":1785728404000,"updated_at":1787315350000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"2.4万亿参数MoE旗舰，编程与办公能力全面跃升，可自主编程十数天交付完整项目。胜任法律、金融、设计等数百种专业任务，一次对话端到端交付生产级成果。原生视觉理解贯穿规划、执行与验证全流程，支持超长文档与长视频的深度语义解析。长程任务中自主规划与闭环迭代，持续进化。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1163,"name":"多模态","slug":"multi-modal","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287910629,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1282,"name":"1M","slug":"context_size_1M","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-ij94b04s1-context_size (1).png","en_name":"1M","color":"blue","description":"上下文大小","href":"","created_at":1781233150090,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28583,"name":"qwen3.8-max","namespace":"moark","title":"qwen3.8-max"},"total_calls":932,"operation_summary":{"service_id":19722,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":12.00000000,"min_output_million_tokens_price":36.00000000,"max_price":0.0000,"max_input_million_tokens_price":12.00000000,"max_output_million_tokens_price":36.00000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":892,"account":0,"account_type":0,"type":"text2text","service_id":19722,"service_type":"serverless","model_id":28583,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":12.00000000,"output_million_tokens_price":36.00000000,"input_million_tokens_price_average":12.00000000,"output_million_tokens_price_average":36.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":1.5,"cache_write_million_tokens_price":12},"version":1,"specified":{"cache_read_million_tokens_price":1.5,"cache_write_million_tokens_price":12}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","type":"number","control":"number","visible":false,"required":false,"description":"temperature","will_be_null":true},{"in":"body","name":"top_p","type":"number","control":"number","visible":false,"required":false,"description":"top_p","will_be_null":true},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","will_be_null":true},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"最大体积：5MB"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1785728538000,"updated_at":1787315457000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[892],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":19831,"ident":"deepseek-v4-flash-0731","type":"serverless","vendor":"","region":"","compute":"","model":28580,"name":"DeepSeek-V4-Flash-0731","remark":"高效轻量化MoE模型，总参284B，激活13B，原生支持百万超长上下文能力。推理速度快、延迟低、调用成本低廉，综合能力均衡，主打高并发、轻量化任务，适合日常对话、内容创作、基础 RAG、批量文案处理等普惠刚需场景。","price":0.0000,"props":{"name":"DeepSeek-V4-Flash-0731","service_type":"serverless"},"inventories":0,"created_at":1785750784000,"updated_at":1787315128000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-00z769y98-deepseek.png","description_html":"高效轻量化MoE模型，总参284B，激活13B，原生支持百万超长上下文能力。推理速度快、延迟低、调用成本低廉，综合能力均衡，主打高并发、轻量化任务，适合日常对话、内容创作、基础 RAG、批量文案处理等普惠刚需场景。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1282,"name":"1M","slug":"context_size_1M","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-ij94b04s1-context_size (1).png","en_name":"1M","color":"blue","description":"上下文大小","href":"","created_at":1781233150090,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28580,"name":"DeepSeek-V4-Flash-0731","namespace":"hf-models","mirror_namespace":"deepseek-ai","mirror_url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731"},"total_calls":1232,"operation_summary":{"service_id":19831,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":3.00000000,"min_output_million_tokens_price":9.00000000,"max_price":0.0000,"max_input_million_tokens_price":3.00000000,"max_output_million_tokens_price":9.00000000,"free_operation_count":2,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":893,"account":0,"account_type":0,"type":"text2text","service_id":19831,"service_type":"serverless","model_id":28580,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":3.00000000,"output_million_tokens_price":9.00000000,"input_million_tokens_price_average":3.00000000,"output_million_tokens_price_average":9.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":0.3,"cache_write_million_tokens_price":3},"version":1,"specified":{"cache_read_million_tokens_price":0.3,"cache_write_million_tokens_price":3}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1785750873000,"updated_at":1787315114000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[893],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":906,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":28580,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216},{"id":906,"account":0,"account_type":0,"type":"text2text","service_id":19831,"service_type":"serverless","model_id":28580,"vendor":"ztzn","region":"ztzn-region","compute":"ztzn/ztzn-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":3.00000000,"output_million_tokens_price":9.00000000,"input_million_tokens_price_average":3.00000000,"output_million_tokens_price_average":9.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":0.3,"cache_write_million_tokens_price":3},"version":1,"specified":{"cache_read_million_tokens_price":0.3,"cache_write_million_tokens_price":3}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1786085231000,"updated_at":1787557543000,"vendor_info":{"id":0,"name":"ztzn","label":"正拓智能","status":"available","ratio":0.70},"region_info":{"name":"ztzn-region","label":"正拓智能默认","props":{}},"compute_info":{"id":"ztzn/ztzn-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"正拓智能","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"ztzn-region","vendor_name":"ztzn","type":"SERVERLESS_API","storage":0},"ids":[906],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":893,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":28580,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":19921,"ident":"Duix-Avatar","type":"serverless","vendor":"","region":"","compute":"","model":0,"name":"Duix-Avatar","remark":"Duix-Avatar 是一个开源 AI 数字人框架，通过语音驱动和视觉生成技术，让虚拟角色实现自然说话、精准口型同步和表情互动。它提供从模型推理到服务部署的完整能力，帮助开发者快速构建智能数字人应用。","price":0.0000,"props":{"name":"Duix-Avatar","service_type":"serverless","resolution_price_factors":{"0-15秒(含)":{"end":15,"start":0,"factor":1},"15-60秒(含)":{"end":60,"start":16,"factor":2},"≤ 720P(含)":{"width":1280,"factor":10,"height":720},"720P-1080P(含)":{"width":1920,"factor":15,"height":1080}}},"inventories":0,"created_at":1782835200000,"updated_at":1785822258000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-11g3myb8s-id-rb71nnf85-olynv1vmvrx.png","description_html":"注意该模型依据视频生成时长与视频分辨率按不同档位全额计费（非分段累加），例如：20秒的1080p视频，全程按 0.3元/秒 计算。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1191,"name":"视频生成","slug":"video-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text_to_video.svg","en_name":"Video Generation","color":"","description":"","href":"","created_at":1739973857876,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1224,"name":"数字人","slug":"Digital_Human","icon":"","en_name":"Digital Human","color":"","description":"","href":"","created_at":1750325639154,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"total_calls":7,"operation_summary":{"service_id":19921,"unit_tag":1264,"min_price":0.0100,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0100,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{"≤ 720P(含)":{"name":"≤ 720P(含)","width":1280,"height":720,"price_factor":10.0,"pixels":921600},"720P-1080P(含)":{"name":"720P-1080P(含)","width":1920,"height":1080,"price_factor":15.0,"pixels":2073600}},"all_vendors":false,"subscription_service":false},"operations":[{"id":894,"account":0,"account_type":0,"type":"audio_video2video","service_id":19921,"service_type":"serverless","model_id":0,"vendor":"gitee","region":"prod-gitee-urc-ip","compute":"gitee/prod-gitee-urc-ip/serverless_api/gpu/large/default","name":"数字人生成","description":"","path":"v1/async/videos/audio-video-to-video","protocol":"http","price":0.0100,"average_price":0.0100,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"form","name":"ref_audio","type":"file","accept":"audio/*","control":"file","category":"main","required":true,"description":"生成视频所需的音频文件，最大支持1分钟的音频，决定了生成的视频长度和计费规则"},{"in":"form","name":"ref_video","type":"file","accept":"video/*","control":"file","category":"main","required":true,"description":"生成视频所需的原始视频，视频中需要包含清晰的人脸，决定了生成视频的分辨率和计费规则"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1785823708000,"updated_at":1785928337000,"vendor_info":{"id":0,"name":"gitee","label":"Gitee","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8n73csd10-image 5.svg"},"region_info":{"name":"prod-gitee-urc-ip","label":"乌鲁木齐","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"gitee/prod-gitee-urc-ip/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"NVIDIA系列","gpu_memory_gb":24,"gpu_performance_rate":"full","region_name":"prod-gitee-urc-ip","vendor_name":"gitee","type":"SERVERLESS_API","storage":0},"ids":[894],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":1264}]},{"service":{"id":21706,"ident":"DeepSeek-V4-Pro-0813","type":"serverless","vendor":"","region":"","compute":"","model":28668,"name":"DeepSeek-V4-Pro-0813","remark":"旗舰级 MoE 大模型，总参1.6T、激活 49B，原生支持百万级超长上下文。依托海量高质量训练数据，具备顶尖数学逻辑、复杂推理、专业代码与长文本深度解析能力，适配高阶科研、复杂办公、深度智能代理等高难度场景。","price":0.0000,"props":{"name":"DeepSeek-V4-Pro-0813","service_type":"serverless"},"inventories":0,"created_at":1786588814000,"updated_at":1787315276000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-00z769y98-deepseek.png","description_html":"旗舰级 MoE 大模型，总参1.6T、激活 49B，原生支持百万级超长上下文。依托海量高质量训练数据，具备顶尖数学逻辑、复杂推理、专业代码与长文本深度解析能力，适配高阶科研、复杂办公、深度智能代理等高难度场景。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1282,"name":"1M","slug":"context_size_1M","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-ij94b04s1-context_size (1).png","en_name":"1M","color":"blue","description":"上下文大小","href":"","created_at":1781233150090,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28668,"name":"DeepSeek-V4-Pro-0813","namespace":"moark","title":"DeepSeek-V4-Pro-0813"},"total_calls":1807,"operation_summary":{"service_id":21706,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":9.00000000,"min_output_million_tokens_price":27.00000000,"max_price":0.0000,"max_input_million_tokens_price":9.00000000,"max_output_million_tokens_price":27.00000000,"free_operation_count":2,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":914,"account":0,"account_type":0,"type":"text2text","service_id":21706,"service_type":"serverless","model_id":28668,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":9.00000000,"output_million_tokens_price":27.00000000,"input_million_tokens_price_average":9.00000000,"output_million_tokens_price_average":27.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":0.9,"cache_write_million_tokens_price":9},"version":1,"specified":{"cache_read_million_tokens_price":0.9,"cache_write_million_tokens_price":9}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1787036463000,"updated_at":1787554837000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[914],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":919,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":28668,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216},{"id":919,"account":0,"account_type":0,"type":"text2text","service_id":21706,"service_type":"serverless","model_id":28668,"vendor":"ztzn","region":"ztzn-region","compute":"ztzn/ztzn-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":9.00000000,"output_million_tokens_price":27.00000000,"input_million_tokens_price_average":9.00000000,"output_million_tokens_price_average":27.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":0.9,"cache_write_million_tokens_price":9},"version":1,"specified":{"cache_read_million_tokens_price":0.9,"cache_write_million_tokens_price":9}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1787105739000,"updated_at":1787556579000,"vendor_info":{"id":0,"name":"ztzn","label":"正拓智能","status":"available","ratio":0.70},"region_info":{"name":"ztzn-region","label":"正拓智能默认","props":{}},"compute_info":{"id":"ztzn/ztzn-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"正拓智能","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"ztzn-region","vendor_name":"ztzn","type":"SERVERLESS_API","storage":0},"ids":[919],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":914,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":28668,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":22137,"ident":"Wan2.2-T2V-A14B","type":"serverless","vendor":"","region":"","compute":"","model":23175,"name":"Wan2.2-T2V-A14B","remark":"Wan2.2-T2V-A14B 是 Wan2.2 系列的文本生成视频（Text-to-Video）模型，采用混合专家架构，可根据文字描述生成具有复杂运动和电影感画面的视频，本服务仅支持 480P 输出。","price":0.0000,"props":{"name":"Wan2.2-T2V-A14B-Diffusers","service_type":"serverless","resolution_price_factors":{}},"inventories":0,"created_at":1782835200000,"updated_at":1787540629000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"\uD83D\uDCD8 概览
Wan2.2-T2V-A14B 是 Wan2.2 系列的文本生成视频模型，可根据自然语言描述生成视频，本服务仅支持 480P 输出。
⚙ 核心技术特性
采用混合专家（MoE）架构，在不同去噪阶段分别处理整体布局与画面细节；支持复杂运动、光影、色调、构图与镜头表现。
⚡ 输出能力
支持 480P 文生视频输出，适用于短视频、创意内容、动画与视觉预演。
\uD83E\uDDEA 应用场景
根据文字描述生成短片；广告与社交媒体素材；影视分镜和创意预演；动画及视频特效内容生成。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1191,"name":"视频生成","slug":"video-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text_to_video.svg","en_name":"Video Generation","color":"","description":"","href":"","created_at":1739973857876,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1198,"name":"720P","slug":"resolution-720p","icon":"","en_name":"720P","color":"","description":"","href":"","created_at":1741688024214,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":23175,"name":"Wan2.2-T2V-A14B-Diffusers","namespace":"hf-models","mirror_namespace":"Wan-AI","mirror_url":"https://huggingface.co/Wan-AI/Wan2.2-T2V-A14B-Diffusers"},"total_calls":1,"operation_summary":{"service_id":22137,"unit_tag":0,"min_price":1.5000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":1.5000,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":912,"account":0,"account_type":0,"type":"text2video","service_id":22137,"service_type":"serverless","model_id":23175,"vendor":"biren-tech","region":"prod-br-sh-wz-ip","compute":"biren-tech/prod-br-sh-wz-ip/serverless_api/gpu/large/contest","name":"文生视频（异步）","description":"根据文本提示词异步生成 480P 视频","path":"v1/async/videos/generations","protocol":"http","price":1.5000,"average_price":1.5000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"num_inference_steps","step":1,"type":"integer","control":"slider","maximum":30,"minimum":1,"required":false,"description":"生成视频时模型执行迭代的次数。数值越大，生成的质量会更高","default_value":10},{"in":"body","name":"negative_prompt","type":"string","control":"text","required":false,"description":"指导模型避免生成所描述的内容。"},{"in":"body","name":"seed","type":"integer","control":"number","required":false,"description":"用于确保随机过程的结果可复现，从而保证实验的一致性。"},{"in":"body","name":"prompt","type":"string","control":"text","samples":[{"value":"年轻的中国女性穿着传统红色古装，红色纱布轻轻遮盖在头上，营造出一种朦胧美感。她轻轻撩起自己头上的红纱盖头，脸上带着温柔的微笑。她的眼睛大而明亮，眉心有一粒小红点，脸部白皙细腻。她黑色长发梳成古典发型，戴着金色和红色相间的耳环。特写镜头，焦段中等，浅景深使人物动作和面部细节更加突出，背景模糊，色调温暖柔和，整体构图细腻精致，具有浓厚的传统文化氛围。"},{"value":"奇幻T台秀，黑发精灵女性穿着白色紧身衣，在T台上快步走来，背后逐渐伸展出一对巨大的黑白羽翼，占满整个画面。姿势优雅，面容冷漠，节奏感走秀，长发飘逸。背景为纯白色，光线从后方打来，产生强烈的圣光背光效果。全身照，天使主题，浅景深，突显翅膀展开过程的细节与震撼，构图对称，视觉冲击力强。"},{"value":"一个 20 岁的东亚年轻女孩从滑滑梯上滑下，然后坐在滑梯口摆了个 pose，双手比耶，眼神俏皮看向镜头。黑色短发扎成双马尾，脸型圆润，身穿蓝色宽松T恤，米色荷叶边短裙，脚穿有星星图案的高帮帆布鞋和蓝色毛线腿套。背景是一个儿童游乐场，滑梯色彩鲜艳，后方有高楼建筑。照片色调明亮，使用大光圈使背景稍微虚化，人物突出，整体构图有趣，充满童趣。"},{"value":"年轻女子在溪流中钓鱼，身穿绿色背心和短裤。戴着浅色渔夫帽和橙色反光太阳镜，手持钓鱼竿，突然有大鱼上钩，开始使劲收杆。背景是绿色的草地和野花，环境清新自然。全身照，标准镜头拍摄，色调自然，浅景深使人物突出。"},{"value":"一位中年男子正在狭窄的泥土坑中工作锄地。男子身穿红色长袖衫，蓝色牛仔裤，手持工具，专注地挖掘泥土。男子皮肤黝黑，短发，脸型瘦长，表情专注。背景中有一些红色砖块，泥土坑周围的泥土显得非常潮湿。照片采用低角度拍摄，突出男子的工作状态，色调偏暖，景深较小，人物和泥土的细节清晰。"}],"category":"main","required":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1786618200000,"updated_at":1787023978000,"vendor_info":{"id":0,"name":"biren-tech","label":"壁仞科技","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-5rdewrmcj-cover-biren.png"},"region_info":{"name":"prod-br-sh-wz-ip","label":"壁仞上海和乌镇","props":{}},"compute_info":{"id":"biren-tech/prod-br-sh-wz-ip/serverless_api/gpu/large/contest","instance_size":"large","instance_type":"contest","price_per_hour":3.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"壁砺系列","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-br-sh-wz-ip","vendor_name":"biren-tech","type":"SERVERLESS_API","storage":0},"ids":[912],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":23103,"ident":"qwen3.8-27b","type":"serverless","vendor":"","region":"","compute":"","model":28713,"name":"Qwen3.8-27B","remark":"Qwen3.8系列27B原生视觉语言Dense模型，模型效果相较3.6-27B重点提升了文本和视觉模态下的编程和办公场景能力，能更可靠地端到端完成复杂任务，输出值得信赖的成果。","price":0.0000,"props":{"name":"Qwen3.8-27B","service_type":"serverless"},"inventories":0,"created_at":1787278673000,"updated_at":1787745443000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3.8系列27B原生视觉语言Dense模型，模型效果相较3.6-27B重点提升了文本和视觉模态下的编程和办公场景能力，能更可靠地端到端完成复杂任务，输出值得信赖的成果。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1282,"name":"1M","slug":"context_size_1M","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-ij94b04s1-context_size (1).png","en_name":"1M","color":"blue","description":"上下文大小","href":"","created_at":1781233150090,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28713,"name":"Qwen3.8-27B","namespace":"moark","title":"Qwen3.8-27B"},"total_calls":112,"operation_summary":{"service_id":23103,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":3.00000000,"min_output_million_tokens_price":12.00000000,"max_price":0.0000,"max_input_million_tokens_price":3.00000000,"max_output_million_tokens_price":12.00000000,"free_operation_count":2,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":920,"account":0,"account_type":0,"type":"text2text","service_id":23103,"service_type":"serverless","model_id":28713,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":3.00000000,"output_million_tokens_price":12.00000000,"input_million_tokens_price_average":3.00000000,"output_million_tokens_price_average":12.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":0.6,"cache_write_million_tokens_price":3},"version":1,"specified":{"cache_read_million_tokens_price":0.6,"cache_write_million_tokens_price":3}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","type":"number","control":"number","visible":false,"required":false,"description":"temperature","will_be_null":true},{"in":"body","name":"top_p","type":"number","control":"number","visible":false,"required":false,"description":"top_p","will_be_null":true},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","will_be_null":true},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"最大体积：5MB"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1787278770000,"updated_at":1787745443000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[920],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":921,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":28713,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216},{"id":921,"account":0,"account_type":0,"type":"text2text","service_id":23103,"service_type":"serverless","model_id":28713,"vendor":"gitee","region":"prod-gitee-urc-ip","compute":"gitee/prod-gitee-urc-ip/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":3.00000000,"output_million_tokens_price":12.00000000,"input_million_tokens_price_average":3.00000000,"output_million_tokens_price_average":12.00000000,"billing_strategy":"standard","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":0.6,"cache_write_million_tokens_price":3},"version":1,"specified":{"cache_read_million_tokens_price":0.6,"cache_write_million_tokens_price":3}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","type":"number","control":"number","visible":false,"required":false,"description":"temperature","will_be_null":true},{"in":"body","name":"top_p","type":"number","control":"number","visible":false,"required":false,"description":"top_p","will_be_null":true},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","will_be_null":true},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"最大体积：5MB"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1787540823000,"updated_at":1787745443000,"vendor_info":{"id":0,"name":"gitee","label":"Gitee","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8n73csd10-image 5.svg"},"region_info":{"name":"prod-gitee-urc-ip","label":"乌鲁木齐","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"gitee/prod-gitee-urc-ip/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"NVIDIA系列","gpu_memory_gb":24,"gpu_performance_rate":"full","region_name":"prod-gitee-urc-ip","vendor_name":"gitee","type":"SERVERLESS_API","storage":0},"ids":[921],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":920,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":28713,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":24974,"ident":"GLM-5.3","type":"serverless","vendor":"","region":"","compute":"","model":28705,"name":"GLM-5.3","remark":"GLM-5.3是智谱迄今编程能力最强的模型，在内部自建体感评测中较GLM-5.2提升50%；网络安全能力：在白盒代码审查与漏洞发现等安全任务中，GLM-5.3的表现持平Mythos 5，展现出面向网络安全防御场景的强大潜力。","price":0.0000,"props":{"name":"GLM-5.3","service_type":"serverless"},"inventories":0,"created_at":1787555031000,"updated_at":1787564283000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-vs79e29q6-avatar-zhipuai.png","description_html":"GLM-5.3是智谱迄今编程能力最强的模型，在内部自建体感评测中较GLM-5.2提升50%；网络安全能力：在白盒代码审查与漏洞发现等安全任务中，GLM-5.3的表现持平Mythos 5，展现出面向网络安全防御场景的强大潜力。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1269,"name":"128K","slug":"context_size_output_128k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-9y385u6wl-id-7pfkojri0-arrow-down.png","en_name":"128K","color":"blue","description":"最大生成长度","href":"","created_at":1768877784722,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1282,"name":"1M","slug":"context_size_1M","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-ij94b04s1-context_size (1).png","en_name":"1M","color":"blue","description":"上下文大小","href":"","created_at":1781233150090,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28705,"name":"GLM-5.3","namespace":"moark","title":"GLM-5.3"},"total_calls":502,"operation_summary":{"service_id":24974,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":8.00000000,"min_output_million_tokens_price":28.00000000,"max_price":0.0000,"max_input_million_tokens_price":8.00000000,"max_output_million_tokens_price":28.00000000,"free_operation_count":2,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":923,"account":0,"account_type":0,"type":"text2text","service_id":24974,"service_type":"serverless","model_id":28705,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":8.00000000,"output_million_tokens_price":28.00000000,"input_million_tokens_price_average":8.00000000,"output_million_tokens_price_average":28.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":2,"cache_write_million_tokens_price":8},"version":1,"specified":{"cache_read_million_tokens_price":2,"cache_write_million_tokens_price":8}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1787555140000,"updated_at":1787556924000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[923],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":924,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":28705,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216},{"id":924,"account":0,"account_type":0,"type":"text2text","service_id":24974,"service_type":"serverless","model_id":28705,"vendor":"ztzn","region":"ztzn-region","compute":"ztzn/ztzn-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":8.00000000,"output_million_tokens_price":28.00000000,"input_million_tokens_price_average":8.00000000,"output_million_tokens_price_average":28.00000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":2,"cache_write_million_tokens_price":8},"version":1,"specified":{"cache_read_million_tokens_price":2,"cache_write_million_tokens_price":8}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1787555272000,"updated_at":1787556924000,"vendor_info":{"id":0,"name":"ztzn","label":"正拓智能","status":"available","ratio":0.70},"region_info":{"name":"ztzn-region","label":"正拓智能默认","props":{}},"compute_info":{"id":"ztzn/ztzn-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"正拓智能","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"ztzn-region","vendor_name":"ztzn","type":"SERVERLESS_API","storage":0},"ids":[924],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":923,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":28705,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":25284,"ident":"qwen3.8-flash","type":"serverless","vendor":"","region":"","compute":"","model":28715,"name":"Qwen3.8-Flash","remark":"Qwen3.8-Flash 是千问最新推出的多模态大模型，兼具强大的理解与生成能力和出色的响应速度。模型原生支持百万级上下文窗口，能够一次性处理超长文档、代码仓库和复杂对话。","price":0.0000,"props":{"name":"Qwen3.8-Flash-Next","service_type":"serverless"},"inventories":0,"created_at":1787810558000,"updated_at":1787810985000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"Qwen3.8-Flash 是千问最新推出的多模态大模型，兼具强大的理解与生成能力和出色的响应速度。模型原生支持百万级上下文窗口，能够一次性处理超长文档、代码仓库和复杂对话。在编程辅助、智能体协作、图文理解等场景中表现尤为出色——无论是自动修复代码、操作桌面应用，还是分析图表与长视频，都能给出准确、高质量的结果。同时兼容 OpenAI 与 Anthropic 主流接口协议，可无缝接入 Claude Code、Codex 等开发者工具，轻松构建高并发应用与智能工作流。凭借优异的性能与极具竞争力的推理成本，Qwen3.8-Flash 是开发者和企业在 AI 应用中兼顾效果与效率的理想选择。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1149,"name":"Function Calling","slug":"function_calling","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1734693747581,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1282,"name":"1M","slug":"context_size_1M","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-ij94b04s1-context_size (1).png","en_name":"1M","color":"blue","description":"上下文大小","href":"","created_at":1781233150090,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28715,"name":"Qwen3.8-Flash-Next","namespace":"hf-models","mirror_namespace":"Qwen","mirror_url":"https://huggingface.co/Qwen/Qwen3.8-Flash-Next"},"total_calls":2802,"operation_summary":{"service_id":25284,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":0.80000000,"min_output_million_tokens_price":2.70000000,"max_price":0.0000,"max_input_million_tokens_price":0.80000000,"max_output_million_tokens_price":2.70000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":930,"account":0,"account_type":0,"type":"text2text","service_id":25284,"service_type":"serverless","model_id":28715,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0.80000000,"output_million_tokens_price":2.70000000,"input_million_tokens_price_average":0.80000000,"output_million_tokens_price_average":2.70000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":0.1,"cache_write_million_tokens_price":0.8},"version":1,"specified":{"cache_read_million_tokens_price":0.1,"cache_write_million_tokens_price":0.8}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","type":"number","control":"number","visible":false,"required":false,"description":"temperature","will_be_null":true},{"in":"body","name":"top_p","type":"number","control":"number","visible":false,"required":false,"description":"top_p","will_be_null":true},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","will_be_null":true},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"},{"in":"body","expr":"messages.content.image_url","name":"messages[0].content[0].image_url","type":"string","accept":".jpeg, .jpg, .png, .webp","control":"file","category":"main","required":false,"max_length":5,"description":"最大体积：5MB"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1787810735000,"updated_at":1787810744000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[930],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":25360,"ident":"wan3.0-video","type":"serverless","vendor":"","region":"","compute":"","model":28711,"name":"Wan-3.0-video","remark":"万相3.0是 all-in-one 视频生成模型，统一支持参考、编辑、复刻、驱动等多种创作功能，四模态全能参考。","price":0.0000,"props":{"name":"wan3.0-video","service_type":"serverless","resolution_price_factors":{"480P":{"width":854,"factor":1,"height":480},"720P":{"width":1280,"factor":2,"height":720},"1080P":{"width":1920,"factor":4,"height":1080}}},"inventories":0,"created_at":1787819446000,"updated_at":1788251733000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"万相3.0是 all-in-one 视频生成模型，统一支持参考、编辑、复刻、驱动等多种创作功能，四模态全能参考，最长支持30秒视频生成，可解析文件、网页及复杂图片。生产级角色一致性保持，音画真实，带来身临其境的视听冲击力。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1191,"name":"视频生成","slug":"video-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text_to_video.svg","en_name":"Video Generation","color":"","description":"","href":"","created_at":1739973857876,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1197,"name":"480P","slug":"resolution-480p","icon":"","en_name":"480P","color":"","description":"","href":"","created_at":1741688008548,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1198,"name":"720P","slug":"resolution-720p","icon":"","en_name":"720P","color":"","description":"","href":"","created_at":1741688024214,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1286,"name":"1080P","slug":"resolution-1080p","icon":"","en_name":"1080P","color":"","description":"","href":"","created_at":1787820533031,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28711,"name":"wan3.0-video","namespace":"moark","title":"wan3.0-video"},"total_calls":1,"operation_summary":{"service_id":25360,"unit_tag":1264,"min_price":0.3000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.3000,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":5},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{"480P":{"name":"480P","width":854,"height":480,"price_factor":1.0,"pixels":409920},"720P":{"name":"720P","width":1280,"height":720,"price_factor":2.0,"pixels":921600},"1080P":{"name":"1080P","width":1920,"height":1080,"price_factor":4.0,"pixels":2073600}},"all_vendors":false,"subscription_service":false},"operations":[{"id":932,"account":0,"account_type":0,"type":"text2video","service_id":25360,"service_type":"serverless","model_id":28711,"vendor":"aliyun","region":"dashscope","compute":"dashscope","name":"参考文件生视频","description":"Wan 3.0 参考文件、网页或音频生成","path":"v1/async/videos/generations","protocol":"http","price":0.3000,"average_price":0.3000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":20,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.negative_prompt","name":"negative_prompt","type":"string","control":"text","required":false,"description":"反向提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"480P","value":"480P"},{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"1080P"},{"in":"body","expr":"$.ratio","name":"ratio","type":"string","control":"select","options":[{"label":"自适应","value":"adaptive"},{"label":"16:9","value":"16:9"},{"label":"4:3","value":"4:3"},{"label":"1:1","value":"1:1"},{"label":"3:4","value":"3:4"},{"label":"9:16","value":"9:16"}],"required":false,"description":"视频宽高比","default_value":"adaptive"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"integer","control":"slider","maximum":30,"minimum":5,"required":false,"description":"输出时长（秒）","default_value":5},{"in":"body","expr":"$.audio","name":"audio","type":"boolean","control":"boolean","required":false,"description":"是否生成音轨","default_value":true},{"in":"body","expr":"$.seed","name":"seed","type":"integer","control":"number","maximum":2147483647,"minimum":0,"visible":false,"required":false,"description":"随机种子"},{"in":"body","expr":"$.prompt_extend","name":"prompt_extend","type":"boolean","control":"boolean","visible":false,"required":false,"description":"是否智能改写提示词","default_value":true},{"in":"body","expr":"$.watermark","name":"watermark","type":"boolean","control":"boolean","visible":false,"required":false,"description":"是否添加水印","default_value":false},{"in":"body","expr":"$.reference_file","name":"reference_file","type":"file","props":{"enableUrl":true},"accept":".docx,.doc,.xlsx,.xls,.pptx,.ppt,.pdf,.txt,.key,.pages,.numbers,.md","adapter":{"aliyun":{"emit":"media","media_type":"file"}},"control":"file","category":"main","required":true,"description":"参考文件"},{"in":"body","expr":"$.reference_link","name":"reference_link","type":"string","adapter":{"aliyun":{"emit":"media","media_type":"link"}},"control":"string","category":"main","required":false,"description":"公开网页 URL"},{"in":"body","expr":"$.reference_audio","name":"reference_audio","type":"file","accept":"audio/*","adapter":{"aliyun":{"emit":"media","media_type":"reference_audio"}},"control":"file","category":"main","encoding":"data-url","required":false,"description":"参考音频"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1787819446000,"updated_at":1787820377000,"ids":[932],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264},{"id":933,"account":0,"account_type":0,"type":"text2video","service_id":25360,"service_type":"serverless","model_id":28711,"vendor":"aliyun","region":"dashscope","compute":"dashscope","name":"参考生视频","description":"Wan 3.0 参考图片或视频生成","path":"v1/async/videos/generations","protocol":"http","price":0.3000,"average_price":0.3000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":20,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.negative_prompt","name":"negative_prompt","type":"string","control":"text","required":false,"description":"反向提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"480P","value":"480P"},{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"1080P"},{"in":"body","expr":"$.ratio","name":"ratio","type":"string","control":"select","options":[{"label":"自适应","value":"adaptive"},{"label":"16:9","value":"16:9"},{"label":"4:3","value":"4:3"},{"label":"1:1","value":"1:1"},{"label":"3:4","value":"3:4"},{"label":"9:16","value":"9:16"}],"required":false,"description":"视频宽高比","default_value":"adaptive"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"integer","control":"slider","maximum":30,"minimum":5,"required":false,"description":"输出时长（秒）","default_value":5},{"in":"body","expr":"$.audio","name":"audio","type":"boolean","control":"boolean","required":false,"description":"是否生成音轨","default_value":true},{"in":"body","expr":"$.seed","name":"seed","type":"integer","control":"number","maximum":2147483647,"minimum":0,"visible":false,"required":false,"description":"随机种子"},{"in":"body","expr":"$.prompt_extend","name":"prompt_extend","type":"boolean","control":"boolean","visible":false,"required":false,"description":"是否智能改写提示词","default_value":true},{"in":"body","expr":"$.watermark","name":"watermark","type":"boolean","control":"boolean","visible":false,"required":false,"description":"是否添加水印","default_value":false},{"in":"body","expr":"$.reference_images","name":"reference_images","type":"array","props":{"textField":false,"imageField":{"name":"","accept":"image/*","max_length":9,"show_array":true}},"accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"reference_image"}},"control":"multimodal","category":"main","encoding":"data-url","required":true,"max_length":9,"min_length":1,"description":"参考图片"},{"in":"body","expr":"$.reference_videos","name":"reference_videos","type":"array","items":{"type":"file","control":"file"},"accept":"video/*","adapter":{"aliyun":{"emit":"media","media_type":"reference_video"}},"control":"file","category":"main","encoding":"data-url","required":false,"max_length":9,"description":"参考视频，可选，最多 9 个"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1787819446000,"updated_at":1787820377000,"ids":[933],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264},{"id":934,"account":0,"account_type":0,"type":"text2video","service_id":25360,"service_type":"serverless","model_id":28711,"vendor":"aliyun","region":"dashscope","compute":"dashscope","name":"首尾帧生视频","description":"Wan 3.0 首尾帧生视频","path":"v1/async/videos/generations","protocol":"http","price":0.3000,"average_price":0.3000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":20,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.negative_prompt","name":"negative_prompt","type":"string","control":"text","required":false,"description":"反向提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"480P","value":"480P"},{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"1080P"},{"in":"body","expr":"$.ratio","name":"ratio","type":"string","control":"select","options":[{"label":"自适应","value":"adaptive"},{"label":"16:9","value":"16:9"},{"label":"4:3","value":"4:3"},{"label":"1:1","value":"1:1"},{"label":"3:4","value":"3:4"},{"label":"9:16","value":"9:16"}],"required":false,"description":"视频宽高比","default_value":"adaptive"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"integer","control":"slider","maximum":30,"minimum":5,"required":false,"description":"输出时长（秒）","default_value":5},{"in":"body","expr":"$.audio","name":"audio","type":"boolean","control":"boolean","required":false,"description":"是否生成音轨","default_value":true},{"in":"body","expr":"$.seed","name":"seed","type":"integer","control":"number","maximum":2147483647,"minimum":0,"visible":false,"required":false,"description":"随机种子"},{"in":"body","expr":"$.prompt_extend","name":"prompt_extend","type":"boolean","control":"boolean","visible":false,"required":false,"description":"是否智能改写提示词","default_value":true},{"in":"body","expr":"$.watermark","name":"watermark","type":"boolean","control":"boolean","visible":false,"required":false,"description":"是否添加水印","default_value":false},{"in":"body","expr":"$.first_frame","name":"first_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"first_frame"}},"control":"file","category":"main","encoding":"data-url","required":true,"description":"首帧图片"},{"in":"body","expr":"$.last_frame","name":"last_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"last_frame"}},"control":"file","category":"main","encoding":"data-url","required":true,"description":"尾帧图片"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1787819446000,"updated_at":1787820377000,"ids":[934],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264},{"id":935,"account":0,"account_type":0,"type":"text2video","service_id":25360,"service_type":"serverless","model_id":28711,"vendor":"aliyun","region":"dashscope","compute":"dashscope","name":"首帧生视频","description":"Wan 3.0 首帧生视频","path":"v1/async/videos/generations","protocol":"http","price":0.3000,"average_price":0.3000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":20,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.negative_prompt","name":"negative_prompt","type":"string","control":"text","required":false,"description":"反向提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"480P","value":"480P"},{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"1080P"},{"in":"body","expr":"$.ratio","name":"ratio","type":"string","control":"select","options":[{"label":"自适应","value":"adaptive"},{"label":"16:9","value":"16:9"},{"label":"4:3","value":"4:3"},{"label":"1:1","value":"1:1"},{"label":"3:4","value":"3:4"},{"label":"9:16","value":"9:16"}],"required":false,"description":"视频宽高比","default_value":"adaptive"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"integer","control":"slider","maximum":30,"minimum":5,"required":false,"description":"输出时长（秒）","default_value":5},{"in":"body","expr":"$.audio","name":"audio","type":"boolean","control":"boolean","required":false,"description":"是否生成音轨","default_value":true},{"in":"body","expr":"$.seed","name":"seed","type":"integer","control":"number","maximum":2147483647,"minimum":0,"visible":false,"required":false,"description":"随机种子"},{"in":"body","expr":"$.prompt_extend","name":"prompt_extend","type":"boolean","control":"boolean","visible":false,"required":false,"description":"是否智能改写提示词","default_value":true},{"in":"body","expr":"$.watermark","name":"watermark","type":"boolean","control":"boolean","visible":false,"required":false,"description":"是否添加水印","default_value":false},{"in":"body","expr":"$.first_frame","name":"first_frame","type":"file","accept":"image/*","adapter":{"aliyun":{"emit":"media","media_type":"first_frame"}},"control":"file","category":"main","encoding":"data-url","required":true,"description":"首帧图片"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1787819446000,"updated_at":1787820377000,"ids":[935],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264},{"id":936,"account":0,"account_type":0,"type":"text2video","service_id":25360,"service_type":"serverless","model_id":28711,"vendor":"aliyun","region":"dashscope","compute":"dashscope","name":"文生视频","description":"Wan 3.0 文生视频","path":"v1/async/videos/generations","protocol":"http","price":0.3000,"average_price":0.3000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":20,"params":[{"in":"body","expr":"$.prompt","name":"prompt","type":"string","control":"text","category":"main","required":true,"description":"视频提示词"},{"in":"body","expr":"$.negative_prompt","name":"negative_prompt","type":"string","control":"text","required":false,"description":"反向提示词"},{"in":"body","expr":"$.resolution","name":"resolution","type":"string","control":"select","options":[{"label":"480P","value":"480P"},{"label":"720P","value":"720P"},{"label":"1080P","value":"1080P"}],"required":false,"description":"视频分辨率","default_value":"1080P"},{"in":"body","expr":"$.ratio","name":"ratio","type":"string","control":"select","options":[{"label":"自适应","value":"adaptive"},{"label":"16:9","value":"16:9"},{"label":"4:3","value":"4:3"},{"label":"1:1","value":"1:1"},{"label":"3:4","value":"3:4"},{"label":"9:16","value":"9:16"}],"required":false,"description":"视频宽高比","default_value":"adaptive"},{"in":"body","expr":"$.duration","name":"duration","step":1,"type":"integer","control":"slider","maximum":30,"minimum":5,"required":false,"description":"输出时长（秒）","default_value":5},{"in":"body","expr":"$.audio","name":"audio","type":"boolean","control":"boolean","required":false,"description":"是否生成音轨","default_value":true},{"in":"body","expr":"$.seed","name":"seed","type":"integer","control":"number","maximum":2147483647,"minimum":0,"visible":false,"required":false,"description":"随机种子"},{"in":"body","expr":"$.prompt_extend","name":"prompt_extend","type":"boolean","control":"boolean","visible":false,"required":false,"description":"是否智能改写提示词","default_value":true},{"in":"body","expr":"$.watermark","name":"watermark","type":"boolean","control":"boolean","visible":false,"required":false,"description":"是否添加水印","default_value":false}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1787819446000,"updated_at":1787820377000,"ids":[936],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["Aliyun"],"visible":true,"unit_tag":1264}]},{"service":{"id":25827,"ident":"GLM-5.3-Flash","type":"serverless","vendor":"","region":"","compute":"","model":28718,"name":"GLM-5.3-Flash","remark":"GLM-5.3-Flash 是 GLM-5 系列首个原生多模态模型，以极致低成本架构实现超越 GLM-5.2 的更强智能。","price":0.0000,"props":{"name":"GLM-5.3-Flash","service_type":"serverless"},"inventories":0,"created_at":1788251397000,"updated_at":1788251397000,"sort_order":0,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-vs79e29q6-avatar-zhipuai.png","description_html":"GLM-5.3-Flash 是 GLM-5 系列首个原生多模态模型，以极致低成本架构实现超越 GLM-5.2 的更强智能。采用 320B 总参 / 18B 激活的稀疏+线性注意力混合架构，注意力计算量与 KV 缓存分别降低 3.01 倍和 4.44 倍。视觉能力原生融入 Coding 循环，可主动观察界面、渲染结果与交互反馈，跨代码、浏览器与图形界面协同完成任务。同时拓展至 Office、金融研究及专业文档场景，能自主拆解目标、调用工具并优化输出，交付 PPTX、PDF、DOCX、XLSX 等高质量成品。
","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":0,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":0,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1269,"name":"128K","slug":"context_size_output_128k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-9y385u6wl-id-7pfkojri0-arrow-down.png","en_name":"128K","color":"blue","description":"最大生成长度","href":"","created_at":1768877784722,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1282,"name":"1M","slug":"context_size_1M","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-ij94b04s1-context_size (1).png","en_name":"1M","color":"blue","description":"上下文大小","href":"","created_at":1781233150090,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1285,"name":"缓存支持","slug":"label_cache_support","icon":"","en_name":"Cache support","color":"","description":"","href":"","created_at":1787311717613,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":28718,"name":"GLM-5.3-Flash","namespace":"hf-models","mirror_namespace":"zai-org","mirror_url":"https://huggingface.co/zai-org/GLM-5.3-Flash"},"total_calls":1953,"operation_summary":{"service_id":25827,"unit_tag":1216,"min_price":0.0000,"min_input_million_tokens_price":0.80000000,"min_output_million_tokens_price":2.80000000,"max_price":0.0000,"max_input_million_tokens_price":0.80000000,"max_output_million_tokens_price":2.80000000,"free_operation_count":1,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"__","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":940,"account":0,"account_type":0,"type":"text2text","service_id":25827,"service_type":"serverless","model_id":28718,"vendor":"bailian","region":"bailian-region","compute":"bailian/bailian-region/serverless_api/gpu/large/default","name":"文本对话","description":"","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0.80000000,"output_million_tokens_price":2.80000000,"input_million_tokens_price_average":0.80000000,"output_million_tokens_price_average":2.80000000,"billing_strategy":"rule_based","billing_policy":{"type":"prompt_cache_tokens","average":{"cache_read_million_tokens_price":0.23,"cache_write_million_tokens_price":0.8},"version":1,"specified":{"cache_read_million_tokens_price":0.23,"cache_write_million_tokens_price":0.8}},"sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":1024},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","step":1,"type":"number","control":"slider","maximum":2,"minimum":-2,"required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","expr":"tools","name":"tools","type":"string","control":"text","visible":false,"required":false,"description":"支持function call"}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1788251541000,"updated_at":1788251593000,"vendor_info":{"id":0,"name":"bailian","label":"阿里云百炼","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-8arffzjl7-bailian.png"},"region_info":{"name":"bailian-region","label":"阿里云百炼默认","props":{}},"compute_info":{"id":"bailian/bailian-region/serverless_api/gpu/large/default","instance_size":"large","instance_type":"default","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":12,"memory_gb":32,"architecture":"","gpu_memory_gb":80,"gpu_performance_rate":"full","region_name":"bailian-region","vendor_name":"bailian","type":"SERVERLESS_API","storage":0},"ids":[940],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["General"],"visible":true,"unit_tag":1216}]},{"service":{"id":114,"ident":"bge-large-zh-v1.5","type":"serverless","vendor":"metax","region":"prod-mx-gz-ip","compute":"metax/prod-mx-gz-ip/serverless_api/gpu/2.xlarge/metax","model":1285,"name":"bge-large-zh-v1.5","remark":"向量维度为 1024。具有多功能、多语言和多粒度特性的文本检索模型，将文本转化为用数值表示的向量形式，用于文本检索、信息推荐、知识挖掘等场景。","price":0.0000,"props":{"name":"bge-large-zh-v1.5","service_type":"serverless"},"inventories":0,"created_at":1718356952000,"updated_at":1760502373000,"sort_order":2,"status":1,"recommend":false,"cover":"https://foruda.gitee.com/images/1725358594967358579/c9fcf1e4_3.png","description_html":"","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1157,"name":"0.5K","slug":"context_size_0.5k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-y4z8cy4z9-context_size.png","en_name":"0.5k","color":"blue","description":"上下文大小","href":"","created_at":1735287675036,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1161,"name":"中文","slug":"zh_language_label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287850060,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1193,"name":"1024 维","slug":"1024-dim","icon":"","en_name":"1024 dim","color":"blue","description":"","href":"","created_at":1740988128644,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1196,"name":"向量化和重排","slug":"embedding-rerank","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-feature-extraction.svg","en_name":"Embedding and Rerank","color":"","description":"","href":"","created_at":1741323767491,"sort_order":14,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":1285,"name":"bge-large-zh-v1.5","namespace":"hf-models","mirror_namespace":"BAAI","mirror_url":"https://huggingface.co/BAAI/bge-large-zh-v1.5"},"total_calls":199339,"operation_summary":{"service_id":114,"unit_tag":0,"min_price":0.0000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0000,"max_input_million_tokens_price":0.07000000,"max_output_million_tokens_price":0.07000000,"free_operation_count":4,"operation_count":4},"available_services":[],"vendors":[],"vendor_region_compute":"metax_prod-mx-gz-ip_metax/prod-mx-gz-ip/serverless_api/gpu/2.xlarge/metax","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":25,"account":0,"account_type":0,"type":"embeddings","service_id":114,"service_type":"serverless","model_id":1285,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"Embedding(小批量)","description":" 与 OpenAI Embedding 接口兼容，将文本向量化，可用于 RAG","path":"v1/embeddings","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"input","type":"string","control":"string","category":"main","required":true,"max_length":25,"min_length":1},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1724289832000,"updated_at":1783564267000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[25],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":344,"account":0,"account_type":0,"type":"embeddings","service_id":0,"model_id":1285,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":792,"account":0,"account_type":0,"type":"embeddings","service_id":114,"service_type":"serverless","model_id":1285,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"Embedding(大批量)","description":" 与 OpenAI Embedding 接口兼容，将文本向量化，可用于 RAG","path":"v1/embeddings","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0.07000000,"output_million_tokens_price":0.07000000,"input_million_tokens_price_average":0.07000000,"output_million_tokens_price_average":0.07000000,"billing_strategy":"standard","billing_policy":{},"sort_order":0,"params":[{"in":"body","name":"input","type":"string","control":"string","category":"main","required":true,"max_length":1000,"min_length":26},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1783564510000,"updated_at":1787909739000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[792],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":804,"account":0,"account_type":0,"type":"embeddings","service_id":0,"model_id":1285,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":344,"account":0,"account_type":0,"type":"embeddings","service_id":114,"service_type":"serverless","model_id":1285,"vendor":"iluvatar","region":"prod-il-sh-if","compute":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","name":"Embedding(小批量)","description":" 与 OpenAI Embedding 接口兼容，将文本向量化，可用于 RAG","path":"v1/embeddings","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":1,"params":[{"in":"body","name":"input","type":"string","control":"string","category":"main","required":true,"max_length":25,"min_length":1},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1750672086000,"updated_at":1783564267000,"vendor_info":{"id":0,"name":"iluvatar","label":"天数智芯","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-lod1uv2gx-Group 62.svg"},"region_info":{"name":"prod-il-sh-if","label":"上海","props":{"products":["APP","ENDPOINT"]}},"compute_info":{"id":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","instance_size":"2.large","instance_type":"mr-v100","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":2,"cpus_num":12,"memory_gb":32,"architecture":"智铠 100","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-il-sh-if","vendor_name":"iluvatar","type":"SERVERLESS_API","storage":0},"ids":[344],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":25,"account":0,"account_type":0,"type":"embeddings","service_id":0,"model_id":1285,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":804,"account":0,"account_type":0,"type":"embeddings","service_id":114,"service_type":"serverless","model_id":1285,"vendor":"iluvatar","region":"prod-il-sh-if","compute":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","name":"Embedding(大批量)","description":" 与 OpenAI Embedding 接口兼容，将文本向量化，可用于 RAG","path":"v1/embeddings","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0.07000000,"output_million_tokens_price":0.07000000,"input_million_tokens_price_average":0.07000000,"output_million_tokens_price_average":0.07000000,"billing_strategy":"standard","billing_policy":{},"sort_order":1,"params":[{"in":"body","name":"input","type":"string","control":"string","category":"main","required":true,"max_length":1000,"min_length":26},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1783564510000,"updated_at":1787909746000,"vendor_info":{"id":0,"name":"iluvatar","label":"天数智芯","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-lod1uv2gx-Group 62.svg"},"region_info":{"name":"prod-il-sh-if","label":"上海","props":{"products":["APP","ENDPOINT"]}},"compute_info":{"id":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","instance_size":"2.large","instance_type":"mr-v100","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":2,"cpus_num":12,"memory_gb":32,"architecture":"智铠 100","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-il-sh-if","vendor_name":"iluvatar","type":"SERVERLESS_API","storage":0},"ids":[804],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":792,"account":0,"account_type":0,"type":"embeddings","service_id":0,"model_id":1285,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":115,"ident":"bge-small-zh-v1.5","type":"serverless","vendor":"metax","region":"prod-mx-gz-ip","compute":"metax/prod-mx-gz-ip/serverless_api/gpu/2.xlarge/metax","model":1390,"name":"bge-small-zh-v1.5","remark":"向量维度为 512。具有多功能、多语言和多粒度特性的文本检索模型，将文本转化为用数值表示的向量形式，用于文本检索、信息推荐、知识挖掘等场景","price":0.0000,"props":{"name":"bge-small-zh-v1.5","service_type":"serverless"},"inventories":0,"created_at":1718358459000,"updated_at":1741323989000,"sort_order":3,"status":1,"recommend":false,"cover":"https://foruda.gitee.com/images/1725358594967358579/c9fcf1e4_3.png","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1157,"name":"0.5K","slug":"context_size_0.5k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-y4z8cy4z9-context_size.png","en_name":"0.5k","color":"blue","description":"上下文大小","href":"","created_at":1735287675036,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1161,"name":"中文","slug":"zh_language_label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287850060,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1195,"name":"512 维","slug":"512-dim","icon":"","en_name":"512 dim","color":"blue","description":"","href":"","created_at":1740988176768,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1196,"name":"向量化和重排","slug":"embedding-rerank","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-feature-extraction.svg","en_name":"Embedding and Rerank","color":"","description":"","href":"","created_at":1741323767491,"sort_order":15,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":1390,"name":"bge-small-zh-v1.5","namespace":"hf-models","mirror_namespace":"BAAI","mirror_url":"https://huggingface.co/BAAI/bge-small-zh-v1.5"},"total_calls":36565,"operation_summary":{"service_id":115,"unit_tag":0,"min_price":0.0000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0000,"max_input_million_tokens_price":0.07000000,"max_output_million_tokens_price":0.07000000,"free_operation_count":2,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"metax_prod-mx-gz-ip_metax/prod-mx-gz-ip/serverless_api/gpu/2.xlarge/metax","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":24,"account":0,"account_type":0,"type":"embeddings","service_id":115,"service_type":"serverless","model_id":1390,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"Embedding(小批量)","description":" 与 OpenAI Embedding 接口兼容，将文本向量化，可用于 RAG","path":"v1/embeddings","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"input","type":"string","control":"string","category":"main","required":true,"max_length":25,"min_length":1}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1724289832000,"updated_at":1783564267000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[24],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":791,"account":0,"account_type":0,"type":"embeddings","service_id":115,"service_type":"serverless","model_id":1390,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"Embedding(大批量)","description":" 与 OpenAI Embedding 接口兼容，将文本向量化，可用于 RAG","path":"v1/embeddings","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0.07000000,"output_million_tokens_price":0.07000000,"input_million_tokens_price_average":0.07000000,"output_million_tokens_price_average":0.07000000,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"input","type":"string","control":"string","category":"main","required":true,"max_length":1000,"min_length":26}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1783564510000,"updated_at":1787558452000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[791],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":1216}]},{"service":{"id":135,"ident":"ChatTTS","type":"serverless","vendor":"vnet","region":"prod-vn-bj-ip","compute":"vnet/prod-vn-bj-ip/serverless_api/gpu/2.xlarge/vnet","model":13714,"name":"ChatTTS","remark":"ChatTTS 是一个专为对话场景设计的语音生成模型，用于对话任务，或对话相关的音视频提供语音。支持中英文，模型由约 10 万小时的中英文数据训练，能够输出自然流畅的语音合成效果。","price":0.0000,"props":{"name":"ChatTTS","service_type":"serverless"},"inventories":0,"created_at":1719374941000,"updated_at":1735288696000,"sort_order":4,"status":1,"recommend":false,"cover":"https://foruda.gitee.com/images/1725358627365890921/ad22ea54_3.png","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":238,"name":"语音和音频合成","slug":"text-to-speech","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-to-speech.svg","en_name":"Text To Speech","color":"","description":"","href":"","created_at":1700443954770,"sort_order":9,"categories":[{"id":13,"parent_id":1,"name":"音频","slug":"audio","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/audio.svg","sort_order":13,"en_name":"Audio"},{"id":19,"parent_id":8,"name":"音频","slug":"dataset-audio","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-audio.svg","sort_order":19,"en_name":"Audio"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1160,"name":"多语言","slug":"multi-language-label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287760476,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":13714,"name":"ChatTTS","namespace":"hf-models","mirror_namespace":"2Noise","mirror_url":"https://huggingface.co/2Noise/ChatTTS"},"total_calls":9356,"operation_summary":{"service_id":135,"unit_tag":0,"min_price":0.0200,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0200,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"vnet_prod-vn-bj-ip_vnet/prod-vn-bj-ip/serverless_api/gpu/2.xlarge/vnet","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":105,"account":0,"account_type":0,"type":"text2speech","service_id":135,"service_type":"serverless","model_id":13714,"vendor":"iluvatar","region":"prod-il-sh-if","compute":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","name":"语音生成","description":"生成文本对应语音，音色自然，中文支持好","path":"v1/audio/speech","protocol":"http","price":0.0200,"average_price":0.0200,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"oral","step":1,"type":"number","control":"slider","maximum":9,"minimum":0,"required":false,"description":"此参数表示口语化程度，取值范围为 [0,9] 数值越小，表达越官方，数值越大，表达越生活化。","default_value":1},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.2},{"in":"body","name":"top_P","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前 p% 的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_K","step":1,"type":"number","control":"slider","maximum":20,"minimum":0,"required":false,"description":"取值范围为 [0,20]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":20},{"in":"body","name":"voice_url","type":"string","control":"text","required":false,"description":".pt 格式的文件链接，用于调整生成语音的音色"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1733992975000,"updated_at":1782999692000,"vendor_info":{"id":0,"name":"iluvatar","label":"天数智芯","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-lod1uv2gx-Group 62.svg"},"region_info":{"name":"prod-il-sh-if","label":"上海","props":{"products":["APP","ENDPOINT"]}},"compute_info":{"id":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","instance_size":"2.large","instance_type":"mr-v100","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":2,"cpus_num":12,"memory_gb":32,"architecture":"智铠 100","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-il-sh-if","vendor_name":"iluvatar","type":"SERVERLESS_API","storage":0},"ids":[105],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":1091,"ident":"bge-m3","type":"serverless","vendor":"vnet","region":"prod-vn-bj-ip","compute":"vnet/prod-vn-bj-ip/serverless_api/gpu/2.xlarge/vnet","model":5558,"name":"bge-m3","remark":"由智源研究院开源的 bge-m3 模型具有多语言、多粒度和多功能的特点，向量维度为 1024，支持100 多种语言，适用于句子到文档的不同文本粒度，最大输入长度为 8192，并集成了稠密、稀疏和多向量检索功能，在多个评测中表现优异。","price":0.0000,"props":{"name":"bge-m3","service_type":"serverless"},"inventories":0,"created_at":1723546832000,"updated_at":1741323996000,"sort_order":5,"status":1,"recommend":true,"cover":"https://foruda.gitee.com/images/1725358594967358579/c9fcf1e4_3.png","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1154,"name":"8K","slug":"context_size_8k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-1e5w7xxiu-context_size.png","en_name":"8K","color":"blue","description":"上下文大小","href":"","created_at":1734693887177,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1160,"name":"多语言","slug":"multi-language-label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287760476,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1193,"name":"1024 维","slug":"1024-dim","icon":"","en_name":"1024 dim","color":"blue","description":"","href":"","created_at":1740988128644,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1196,"name":"向量化和重排","slug":"embedding-rerank","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-feature-extraction.svg","en_name":"Embedding and Rerank","color":"","description":"","href":"","created_at":1741323767491,"sort_order":12,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":5558,"name":"bge-m3","namespace":"hf-models","mirror_namespace":"BAAI","mirror_url":"https://huggingface.co/BAAI/bge-m3"},"total_calls":2087097,"operation_summary":{"service_id":1091,"unit_tag":0,"min_price":0.0000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0000,"max_input_million_tokens_price":0.07000000,"max_output_million_tokens_price":0.07000000,"free_operation_count":4,"operation_count":4},"available_services":[],"vendors":[],"vendor_region_compute":"vnet_prod-vn-bj-ip_vnet/prod-vn-bj-ip/serverless_api/gpu/2.xlarge/vnet","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":260,"account":0,"account_type":0,"type":"embeddings","service_id":1091,"service_type":"serverless","model_id":5558,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"Embedding(小批量)","description":"具有多功能、多语言和多粒度特性的文本检索模型，将文本转化为用数值表示的向量形式，用于文本检索、信息推荐、知识挖掘等场景","path":"v1/embeddings","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"input","type":"string","control":"string","category":"main","required":true,"max_length":25,"min_length":1},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1745315239000,"updated_at":1783564267000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[260],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":71,"account":0,"account_type":0,"type":"embeddings","service_id":0,"model_id":5558,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":797,"account":0,"account_type":0,"type":"embeddings","service_id":1091,"service_type":"serverless","model_id":5558,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"Embedding(大批量)","description":"具有多功能、多语言和多粒度特性的文本检索模型，将文本转化为用数值表示的向量形式，用于文本检索、信息推荐、知识挖掘等场景","path":"v1/embeddings","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0.07000000,"output_million_tokens_price":0.07000000,"input_million_tokens_price_average":0.07000000,"output_million_tokens_price_average":0.07000000,"billing_strategy":"standard","billing_policy":{},"sort_order":0,"params":[{"in":"body","name":"input","type":"string","control":"string","category":"main","required":true,"max_length":1000,"min_length":26},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1783564510000,"updated_at":1787909660000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[797],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":793,"account":0,"account_type":0,"type":"embeddings","service_id":0,"model_id":5558,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":71,"account":0,"account_type":0,"type":"embeddings","service_id":1091,"service_type":"serverless","model_id":5558,"vendor":"iluvatar","region":"prod-il-sh-if","compute":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","name":"Embedding(小批量)","description":"具有多功能、多语言和多粒度特性的文本检索模型，将文本转化为用数值表示的向量形式，用于文本检索、信息推荐、知识挖掘等场景","path":"v1/embeddings","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":1,"params":[{"in":"body","name":"input","type":"string","control":"string","category":"main","required":true,"max_length":25,"min_length":1},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1731311131000,"updated_at":1783564267000,"vendor_info":{"id":0,"name":"iluvatar","label":"天数智芯","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-lod1uv2gx-Group 62.svg"},"region_info":{"name":"prod-il-sh-if","label":"上海","props":{"products":["APP","ENDPOINT"]}},"compute_info":{"id":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","instance_size":"2.large","instance_type":"mr-v100","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":2,"cpus_num":12,"memory_gb":32,"architecture":"智铠 100","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-il-sh-if","vendor_name":"iluvatar","type":"SERVERLESS_API","storage":0},"ids":[71],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":260,"account":0,"account_type":0,"type":"embeddings","service_id":0,"model_id":5558,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":793,"account":0,"account_type":0,"type":"embeddings","service_id":1091,"service_type":"serverless","model_id":5558,"vendor":"iluvatar","region":"prod-il-sh-if","compute":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","name":"Embedding(大批量)","description":"具有多功能、多语言和多粒度特性的文本检索模型，将文本转化为用数值表示的向量形式，用于文本检索、信息推荐、知识挖掘等场景","path":"v1/embeddings","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0.07000000,"output_million_tokens_price":0.07000000,"input_million_tokens_price_average":0.07000000,"output_million_tokens_price_average":0.07000000,"billing_strategy":"standard","billing_policy":{},"sort_order":1,"params":[{"in":"body","name":"input","type":"string","control":"string","category":"main","required":true,"max_length":1000,"min_length":26},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":false,"api_format":"OPEN_AI","created_at":1783564510000,"updated_at":1787909678000,"vendor_info":{"id":0,"name":"iluvatar","label":"天数智芯","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-lod1uv2gx-Group 62.svg"},"region_info":{"name":"prod-il-sh-if","label":"上海","props":{"products":["APP","ENDPOINT"]}},"compute_info":{"id":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","instance_size":"2.large","instance_type":"mr-v100","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":2,"cpus_num":12,"memory_gb":32,"architecture":"智铠 100","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-il-sh-if","vendor_name":"iluvatar","type":"SERVERLESS_API","storage":0},"ids":[793],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":797,"account":0,"account_type":0,"type":"embeddings","service_id":0,"model_id":5558,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":1090,"ident":"bge-reranker-v2-m3","type":"serverless","vendor":"vnet","region":"prod-vn-bj-ip","compute":"vnet/prod-vn-bj-ip/serverless_api/gpu/2.xlarge/vnet","model":13092,"name":"bge-reranker-v2-m3","remark":"bge-reranker-v2-m3 是一款支持多语言、高精度语义匹配和高效检索优化的文本重排序模型。","price":0.0000,"props":{"name":"bge-reranker-v2-m3","service_type":"serverless"},"inventories":0,"created_at":1723546823000,"updated_at":1741324001000,"sort_order":6,"status":1,"recommend":false,"cover":"https://foruda.gitee.com/images/1725358594967358579/c9fcf1e4_3.png","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1154,"name":"8K","slug":"context_size_8k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-1e5w7xxiu-context_size.png","en_name":"8K","color":"blue","description":"上下文大小","href":"","created_at":1734693887177,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1160,"name":"多语言","slug":"multi-language-label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287760476,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1196,"name":"向量化和重排","slug":"embedding-rerank","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-feature-extraction.svg","en_name":"Embedding and Rerank","color":"","description":"","href":"","created_at":1741323767491,"sort_order":13,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]}],"model_info":{"id":13092,"name":"bge-reranker-v2-m3","namespace":"hf-models","mirror_namespace":"BAAI","mirror_url":"https://huggingface.co/BAAI/bge-reranker-v2-m3"},"total_calls":246348,"operation_summary":{"service_id":1090,"unit_tag":0,"min_price":0.0000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0000,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":4,"operation_count":4},"available_services":[],"vendors":[],"vendor_region_compute":"vnet_prod-vn-bj-ip_vnet/prod-vn-bj-ip/serverless_api/gpu/2.xlarge/vnet","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":190,"account":0,"account_type":0,"type":"sentence_similarity","service_id":1090,"service_type":"serverless","model_id":13092,"vendor":"iluvatar","region":"prod-il-sh-if","compute":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","name":"句子相似度对比","description":"对比多个句子和原句的相似度，用于 Rerank 重排","path":"v1/sentence-similarity","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1732647932000,"updated_at":1775273954000,"vendor_info":{"id":0,"name":"iluvatar","label":"天数智芯","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-lod1uv2gx-Group 62.svg"},"region_info":{"name":"prod-il-sh-if","label":"上海","props":{"products":["APP","ENDPOINT"]}},"compute_info":{"id":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","instance_size":"2.large","instance_type":"mr-v100","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":2,"cpus_num":12,"memory_gb":32,"architecture":"智铠 100","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-il-sh-if","vendor_name":"iluvatar","type":"SERVERLESS_API","storage":0},"ids":[190],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":625,"account":0,"account_type":0,"type":"sentence_similarity","service_id":0,"model_id":13092,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":625,"account":0,"account_type":0,"type":"sentence_similarity","service_id":1090,"service_type":"serverless","model_id":13092,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"句子相似度对比","description":"对比多个句子和原句的相似度，用于 Rerank 重排","path":"v1/sentence-similarity","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1770192369000,"updated_at":1773910953000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[625],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":190,"account":0,"account_type":0,"type":"sentence_similarity","service_id":0,"model_id":13092,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":626,"account":0,"account_type":0,"type":"rerank","service_id":1090,"service_type":"serverless","model_id":13092,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"句子重排","description":"完成句子相似度对比，并支持重新排序和筛选","path":"v1/rerank","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"query","type":"string","control":"string","samples":[{"label":"Python 读取 CVS 文件","value":"How to read a CSV file in Python?"}],"category":"main","required":true},{"in":"body","name":"documents","type":"array","control":"multimodal","samples":[{"label":"Python 读取 CVS 文件","value":[{"type":"text","value":"Use pandas: `import pandas as pd; df = pd.read_csv('data.csv')`"},{"type":"text","value":"You can read CSV files with numpy.loadtxt()"},{"type":"text","value":"To write JSON files, use json.dump() in Python"},{"type":"text","value":"CSV means Comma Separated Values. Python files can be opened using read() method."}]}],"category":"main","required":true,"max_length":25,"min_length":1},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1770192395000,"updated_at":1773910970000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[626],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":187,"account":0,"account_type":0,"type":"rerank","service_id":0,"model_id":13092,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":187,"account":0,"account_type":0,"type":"rerank","service_id":1090,"service_type":"serverless","model_id":13092,"vendor":"iluvatar","region":"prod-il-sh-if","compute":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","name":"句子重排","description":"完成句子重新排序和筛选","path":"v1/rerank","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":2,"params":[{"in":"body","name":"query","type":"string","control":"string","samples":[{"label":"Python 读取 CVS 文件","value":"How to read a CSV file in Python?"}],"category":"main","required":true},{"in":"body","name":"documents","type":"array","control":"multimodal","samples":[{"label":"Python 读取 CVS 文件","value":[{"type":"text","value":"Use pandas: `import pandas as pd; df = pd.read_csv('data.csv')`"},{"type":"text","value":"You can read CSV files with numpy.loadtxt()"},{"type":"text","value":"To write JSON files, use json.dump() in Python"},{"type":"text","value":"CSV means Comma Separated Values. Python files can be opened using read() method."}]}],"category":"main","required":true,"max_length":25,"min_length":1},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1732647932000,"updated_at":1775273954000,"vendor_info":{"id":0,"name":"iluvatar","label":"天数智芯","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-lod1uv2gx-Group 62.svg"},"region_info":{"name":"prod-il-sh-if","label":"上海","props":{"products":["APP","ENDPOINT"]}},"compute_info":{"id":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","instance_size":"2.large","instance_type":"mr-v100","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":2,"cpus_num":12,"memory_gb":32,"architecture":"智铠 100","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-il-sh-if","vendor_name":"iluvatar","type":"SERVERLESS_API","storage":0},"ids":[187],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":626,"account":0,"account_type":0,"type":"rerank","service_id":0,"model_id":13092,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":144,"ident":"deepseek-coder-33B-instruct","type":"serverless","vendor":"vnet","region":"prod-vn-bj-ip","compute":"vnet/prod-vn-bj-ip/serverless_api/gpu/2.xlarge/vnet","model":16117,"name":"deepseek-coder-33B-instruct","remark":"DeepSeek Coder 33B 是一个代码语言模型， 基于 2 万亿数据训练而成，其中 87% 为代码， 13% 为中英文语言。模型引入 16K 窗口大小和填空任务，提供项目级别的代码补全和片段填充功能。","price":0.0000,"props":{"name":"deepseek-coder-33b-instruct","service_type":"serverless"},"inventories":0,"created_at":1720398836000,"updated_at":1750934871000,"sort_order":7,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-00z769y98-deepseek.png","description_html":"","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1146,"name":"AI 编程模型","slug":"code-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-code-generation.svg","en_name":"Code Generation","color":"","description":"","href":"","created_at":1724824410309,"sort_order":40,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1154,"name":"8K","slug":"context_size_8k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-1e5w7xxiu-context_size.png","en_name":"8K","color":"blue","description":"上下文大小","href":"","created_at":1734693887177,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":16117,"name":"deepseek-coder-33b-instruct","namespace":"hf-models","mirror_namespace":"deepseek-ai","mirror_url":"https://huggingface.co/deepseek-ai/deepseek-coder-33b-instruct"},"total_calls":17124,"operation_summary":{"service_id":144,"unit_tag":0,"min_price":0.0200,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0200,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"vnet_prod-vn-bj-ip_vnet/prod-vn-bj-ip/serverless_api/gpu/2.xlarge/vnet","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":97,"account":0,"account_type":0,"type":"text2text","service_id":144,"service_type":"serverless","model_id":16117,"vendor":"iluvatar","region":"prod-il-sh-if","compute":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","name":"文本对话","description":"与 OpenAI 对话接口兼容，上下文大小为 8K，擅长代码任务","path":"v1/chat/completions","protocol":"http","price":0.0200,"average_price":0.0200,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":512},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","type":"number","control":"number","required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低"},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"body","name":"messages","type":"array","control":"array","samples":[{"group":"代码助手","label":"代码改写","value":[{"role":"user","content":"下面这段的代码的效率很低，且没有处理边界情况。请先解释这段代码的问题与解决方法，然后进行优化：\n```\ndef fib(n):\n if n <= 2:\n return n\n return fib(n-1) + fib(n-2)\n```\n"}]},{"group":"代码助手","label":"代码解释","value":[{"role":"user","content":"请解释下面这段代码的逻辑，并说明完成了什么功能：\n```\n// weight数组的大小 就是物品个数\nfor(int i = 1; i < weight.size(); i++) { // 遍历物品\n for(int j = 0; j <= bagweight; j++) { // 遍历背包容量\n if (j < weight[i]) dp[i][j] = dp[i - 1][j];\n else dp[i][j] = max(dp[i - 1][j], dp[i - 1][j - weight[i]] + value[i]);\n }\n}\n```\n"}]},{"group":"代码助手","label":"代码生成","value":[{"role":"user","content":"请帮我用 HTML 生成一个五子棋游戏，所有代码都保存在一个 HTML 中。"}]},{"group":"代码助手","label":"结构化输出","value":[{"role":"system","content":"用户将提供给你一段新闻内容，请你分析新闻内容，并提取其中的关键信息，以 JSON 的形式输出，输出的 JSON 需遵守以下的格式：\n\n{\n \"entiry\": <新闻实体>,\n \"time\": <新闻时间，格式为 YYYY-mm-dd HH:MM:SS，没有请填 null>,\n \"summary\": <新闻内容总结>\n}"},{"role":"user","content":"8月31日，一枚猎鹰9号运载火箭于美国东部时间凌晨3时43分从美国佛罗里达州卡纳维拉尔角发射升空，将21颗星链卫星（Starlink）送入轨道。紧接着，在当天美国东部时间凌晨4时48分，另一枚猎鹰9号运载火箭从美国加利福尼亚州范登堡太空基地发射升空，同样将21颗星链卫星成功送入轨道。两次发射间隔65分钟创猎鹰9号运载火箭最短发射间隔纪录。\n\n美国联邦航空管理局于8月30日表示，尽管对太空探索技术公司的调查仍在进行，但已允许其猎鹰9号运载火箭恢复发射。目前，双方并未透露8月28日助推器着陆失败事故的详细信息。尽管发射已恢复，但原计划进行五天太空活动的“北极星黎明”（Polaris Dawn）任务却被推迟。美国太空探索技术公司为该任务正在积极筹备，等待美国联邦航空管理局的最终批准后尽快进行发射。"}]},{"group":"文本创作","label":"散文写作","value":[{"role":"user","content":"以孤独的夜行者为题写一篇750字的散文，描绘一个人在城市中夜晚漫无目的行走的心情与所见所感，以及夜的寂静给予的独特感悟。"}]},{"group":"文本创作","label":"诗歌创作","value":[{"role":"user","content":"模仿李白的风格写一首七律.飞机"}]},{"group":"文本创作","label":"文案大纲生成","value":[{"role":"system","content":"你是一位文本大纲生成专家，擅长根据用户的需求创建一个有条理且易于扩展成完整文章的大纲，你拥有强大的主题分析能力，能准确提取关键信息和核心要点。具备丰富的文案写作知识储备，熟悉各种文体和题材的文案大纲构建方法。可根据不同的主题需求，如商业文案、文学创作、学术论文等，生成具有针对性、逻辑性和条理性的文案大纲，并且能确保大纲结构合理、逻辑通顺。该大纲应该包含以下部分：\n引言：介绍主题背景，阐述撰写目的，并吸引读者兴趣。\n主体部分：第一段落：详细说明第一个关键点或论据，支持观点并引用相关数据或案例。\n第二段落：深入探讨第二个重点，继续论证或展开叙述，保持内容的连贯性和深度。\n第三段落：如果有必要，进一步讨论其他重要方面，或者提供不同的视角和证据。\n结论：总结所有要点，重申主要观点，并给出有力的结尾陈述，可以是呼吁行动、提出展望或其他形式的收尾。\n创意性标题：为文章构思一个引人注目的标题，确保它既反映了文章的核心内容又能激发读者的好奇心。"},{"role":"user","content":"请帮我生成“中国农业情况”这篇文章的大纲"}]},{"group":"文本创作","label":"宣传标语生成","value":[{"role":"system","content":"你是一个宣传标语专家，请根据用户需求设计一个独具创意且引人注目的宣传标语，需结合该产品/活动的核心价值和特点，同时融入新颖的表达方式或视角。请确保标语能够激发潜在客户的兴趣，并能留下深刻印象，可以考虑采用比喻、双关或其他修辞手法来增强语言的表现力。标语应简洁明了，需要朗朗上口，易于理解和记忆，一定要押韵，不要太过书面化。只输出宣传标语，不用解释。"},{"role":"user","content":"请生成”希腊酸奶“的宣传标语"}]},{"group":"角色扮演","label":"情景续写","value":[{"role":"user","content":"假设诸葛亮死后在地府遇到了刘备，请模拟两个人展开一段对话。"}]},{"group":"角色扮演","label":"自定义人设","value":[{"role":"system","content":"请你扮演一个刚从美国留学回国的人，说话时候会故意中文夹杂部分英文单词，显得非常fancy，对话中总是带有很强的优越感。"},{"role":"user","content":"美国的饮食还习惯么。"}]},{"group":"其他","label":"模型提示词生成","value":[{"role":"system","content":"你是一位大模型提示词生成专家，请根据用户的需求编写一个智能助手的提示词，来指导大模型进行内容生成，要求：\n1. 以 Markdown 格式输出\n2. 贴合用户需求，描述智能助手的定位、能力、知识储备\n3. 提示词应清晰、精确、易于理解，在保持质量的同时，尽可能简洁\n4. 只输出提示词，不要输出多余解释"},{"role":"user","content":"请帮我生成一个“Linux 助手”的提示词"}]},{"group":"其他","label":"中英翻译专家","value":[{"role":"system","content":"你是一个中英文翻译专家，将用户输入的中文翻译成英文，或将用户输入的英文翻译成中文。对于非中文内容，它将提供中文翻译结果。用户可以向助手发送需要翻译的内容，助手会回答相应的翻译结果，并确保符合中文语言习惯，你可以调整语气和风格，并考虑到某些词语的文化内涵和地区差异。同时作为翻译家，需将原文翻译成具有信达雅标准的译文。\"信\" 即忠实于原文的内容与意图；\"达\" 意味着译文应通顺易懂，表达清晰；\"雅\" 则追求译文的文化审美和语言的优美。目标是创作出既忠于原作精神，又符合目标语言文化和读者审美的翻译。"},{"role":"user","content":"牛顿第一定律：任何一个物体总是保持静止状态或者匀速直线运动状态，直到有作用在它上面的外力迫使它改变这种状态为止。 如果作用在物体上的合力为零，则物体保持匀速直线运动。 即物体的速度保持不变且加速度为零。"}]}],"visible":true,"category":"main","required":false}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1733222053000,"updated_at":1775273954000,"vendor_info":{"id":0,"name":"iluvatar","label":"天数智芯","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-lod1uv2gx-Group 62.svg"},"region_info":{"name":"prod-il-sh-if","label":"上海","props":{"products":["APP","ENDPOINT"]}},"compute_info":{"id":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","instance_size":"2.large","instance_type":"mr-v100","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":2,"cpus_num":12,"memory_gb":32,"architecture":"智铠 100","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-il-sh-if","vendor_name":"iluvatar","type":"SERVERLESS_API","storage":0},"ids":[97],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":98,"account":0,"account_type":0,"type":"completions","service_id":144,"service_type":"serverless","model_id":16117,"vendor":"iluvatar","region":"prod-il-sh-if","compute":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","name":"文本补全","description":"与 OpenAI 补全接口兼容，上下文大小为 8K，擅长代码补全任务","path":"v1/completions","protocol":"http","price":0.0200,"average_price":0.0200,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":1,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":512},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"frequency_penalty","type":"number","control":"number","required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户。"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1733222133000,"updated_at":1775273954000,"vendor_info":{"id":0,"name":"iluvatar","label":"天数智芯","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-lod1uv2gx-Group 62.svg"},"region_info":{"name":"prod-il-sh-if","label":"上海","props":{"products":["APP","ENDPOINT"]}},"compute_info":{"id":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","instance_size":"2.large","instance_type":"mr-v100","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":2,"cpus_num":12,"memory_gb":32,"architecture":"智铠 100","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-il-sh-if","vendor_name":"iluvatar","type":"SERVERLESS_API","storage":0},"ids":[98],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":1512,"ident":"flux-1-schnell","type":"serverless","vendor":"metax","region":"prod-mx-gz-ip","compute":"metax/prod-mx-gz-ip/serverless_api/gpu/2.xlarge/metax","model":17019,"name":"FLUX.1-schnell","remark":"由 Black Forest Labs 开发的 120 亿参数文生图模型，采用潜在对抗扩散蒸馏技术，能够在 1 到 4 步内生成高质量图像。该模型性能媲美闭源替代品，并在 Apache-2.0 许可证下发布，适用于个人、科研和商业用途。","price":0.0000,"props":{"name":"FLUX.1-schnell","service_type":"serverless"},"inventories":0,"created_at":1725354000000,"updated_at":1765264665000,"sort_order":9,"status":1,"recommend":false,"cover":"https://foruda.gitee.com/images/1725353234915042863/92549b63_3.png","description_html":"","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1148,"name":"图像生成与处理","slug":"image_generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-to-image.svg","en_name":"Image Generation","color":"","description":"","href":"","created_at":1734660944141,"sort_order":3,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1159,"name":"1536*1536","slug":"1536_1536_label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287722805,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1161,"name":"中文","slug":"zh_language_label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287850060,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1162,"name":"LoRA","slug":"lora","icon":"","en_name":"LoRA","color":"blue","description":"","href":"","created_at":1735287887794,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1172,"name":"人脸迁移","slug":"face-migration-label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735647662843,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1174,"name":"英文","slug":"en_language_label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735647874656,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":17019,"name":"FLUX.1-schnell","namespace":"hf-models","mirror_namespace":"black-forest-labs","mirror_url":"https://huggingface.co/black-forest-labs/FLUX.1-schnell"},"total_calls":511115,"operation_summary":{"service_id":1512,"unit_tag":1227,"min_price":0.0050,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0120,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":7},"available_services":[],"vendors":[],"vendor_region_compute":"metax_prod-mx-gz-ip_metax/prod-mx-gz-ip/serverless_api/gpu/2.xlarge/metax","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":62,"account":0,"account_type":0,"type":"text2image","service_id":1512,"service_type":"serverless","model_id":17019,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"文生图(1024*1024)","description":"支持生成 1024*1024 内分辨率的图像，支持动态尺寸","path":"v1/images/generations","protocol":"http","price":0.0050,"average_price":0.0050,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"size","type":"string","control":"select","options":[{"label":"1:1 (256*256)","value":"256x256"},{"label":"1:1 (512*512)","value":"512x512"},{"label":"1:1 (1024*1024)","value":"1024x1024"},{"label":"4:3 (1024*768)","value":"1024x768"},{"label":"3:4 (768*1024)","value":"768x1024"},{"label":"16:9 (1024*576)","value":"1024x576"},{"label":"9:16 (576*1024)","value":"576x1024"},{"label":"3:2 (1024*640)","value":"1024x640"},{"label":"2:3(640*1024)","value":"640x1024"}],"required":false,"description":"图片尺寸","default_value":"1024x1024"},{"in":"body","name":"guidance_scale","step":0.1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"调节模型在生成图像时对输入提示的遵从程度，数值越大，遵从越严格","default_value":7.5},{"in":"body","name":"seed","type":"integer","control":"number","required":false,"description":"用于确保随机过程的结果可复现，从而保证实验的一致性。","default_value":42},{"in":"body","name":"prompt","type":"string","control":"string","samples":[{"label":"机器人","value":"机器人坐在露天草地上，正在画布上画画"},{"label":"女孩","value":"一个可爱的日本女孩，正面全身，坐在汽车左侧乘客座位上，系着安全带，用索尼 α7R IV 和电影灯光拍摄，超现实主义摄影"},{"label":"卡通角色","value":"身穿和服的卡通武士插图，正面，全身，带花朵图案， 2D 对称构图"}],"category":"main","required":true},{"in":"body","name":"lora_weights","type":"array","control":"array","required":false,"max_length":4,"min_length":0,"description":"lora权重配置"},{"in":"body","name":"lora_weights.url","type":"string","control":"text","required":false,"description":"lora模型权重的url，若为空则视为基础sdxl模型的推理。"},{"in":"body","name":"lora_weights.weight","type":"number","control":"number","required":false,"description":"若存在多个lora，此处填写长度须与lora数量相同，否则报错。这些数值用于不同lora权重之间的加权融合。","default_value":1},{"in":"body","name":"lora_scale","type":"number","control":"number","required":false,"description":"融合的lora权重对基础模型的影响程度。"},{"in":"body","name":"num_images_per_prompt","type":"integer","control":"number","maximum":4,"minimum":1,"required":false,"description":"生成图片张数，计费将根据生成张数*单价计费。"},{"in":"body","name":"width","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的宽度，最大为1024，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"height","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的高度， 最大为1024，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"response_format","type":"string","control":"select","options":[{"label":"b64_json","value":"b64_json"},{"label":"url","value":"url"}],"required":false,"description":"生成图像时返回的格式。必须为 url 或 b64_json 之一。"},{"in":"body","name":"negative_prompt","type":"string","control":"text","required":false,"description":"负向提示词"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1731311131000,"updated_at":1776823984000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[62],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":1227},{"id":63,"account":0,"account_type":0,"type":"text2image","service_id":1512,"service_type":"serverless","model_id":17019,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"文生图(1536*1536)","description":"支持生成 1536*1536 内分辨率的图像，支持动态尺寸","path":"v1/images/generations","protocol":"http","price":0.0100,"average_price":0.0100,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":1,"params":[{"in":"body","name":"size","type":"string","control":"select","options":[{"label":"1:1 (1536*1536)","value":"1536x1536"},{"label":"4:3 (1536*1152)","value":"1536x1152"},{"label":"3:4 (1152*1536)","value":"1152x1536"},{"label":"16:9 (1536*864)","value":"1536x864"},{"label":"9:16 (864*1536)","value":"864x1536"},{"label":"3:2 (1536*1024)","value":"1536x1024"},{"label":"2:3 (1024*1536)","value":"1024x1536"}],"required":false,"description":"图片尺寸","default_value":"1536x1536"},{"in":"body","name":"guidance_scale","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"调节模型在生成图像时对输入提示的遵从程度，数值越大，遵从越严格","default_value":7.5},{"in":"body","name":"seed","type":"integer","control":"number","required":false,"description":"用于确保随机过程的结果可复现，从而保证实验的一致性。","default_value":42},{"in":"body","name":"prompt","type":"string","control":"string","samples":[{"label":"机器人","value":"机器人坐在露天草地上，正在画布上画画"},{"label":"女孩","value":"一个可爱的日本女孩，正面全身，坐在汽车左侧乘客座位上，系着安全带，用索尼 α7R IV 和电影灯光拍摄，超现实主义摄影"},{"label":"卡通角色","value":"身穿和服的卡通武士插图，正面，全身，带花朵图案， 2D 对称构图"}],"category":"main","required":true},{"in":"body","name":"lora_weights","type":"array","control":"array","required":false,"max_length":4,"min_length":0,"description":"lora权重配置"},{"in":"body","name":"lora_weights.url","type":"string","control":"text","required":false,"description":"lora模型权重的url，若为空则视为基础sdxl模型的推理。"},{"in":"body","name":"lora_weights.weight","type":"number","control":"number","required":false,"description":"若存在多个lora，此处填写长度须与lora数量相同，否则报错。这些数值用于不同lora权重之间的加权融合。","default_value":1},{"in":"body","name":"lora_scale","type":"number","control":"number","required":false,"description":"融合的lora权重对基础模型的影响程度。"},{"in":"body","name":"num_images_per_prompt","type":"integer","control":"number","maximum":4,"minimum":1,"required":false,"description":"生成图片张数，计费将根据生成张数*单价计费。"},{"in":"body","name":"width","step":8,"type":"integer","control":"number","maximum":1536,"minimum":1024,"required":false,"description":"生成图片的宽度，最大为1536，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"height","step":8,"type":"integer","control":"number","maximum":1536,"minimum":1024,"required":false,"description":"生成图片的高度， 最大为1536，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"response_format","type":"string","control":"select","options":[{"label":"b64_json","value":"b64_json"},{"label":"url","value":"url"}],"required":false,"description":"生成图像时返回的格式。必须为 url 或 b64_json 之一。"},{"in":"body","name":"negative_prompt","type":"string","control":"text","required":false,"description":"负向提示词"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1731311131000,"updated_at":1774611824000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[63],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":1227},{"id":377,"account":0,"account_type":0,"type":"text2image","service_id":1512,"service_type":"serverless","model_id":17019,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"文生图(2048*2048)","description":"","path":"v1/images/generations","protocol":"http","price":0.0120,"average_price":0.0120,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":2,"params":[{"in":"body","name":"size","type":"string","control":"select","options":[{"label":"1:1 (2048×2048)","value":"2048x2048"},{"label":"4:3 (2048×1536)","value":"2048x1536"},{"label":"3:4 (1536×2048)","value":"1536x2048"},{"label":"3:2 (2048×1360)","value":"2048x1360"},{"label":"2:3 (1360×2048)","value":"1360x2048"},{"label":"16:9 (2048×1152)","value":"2048x1152"},{"label":"9:16 (1152×2048)","value":"1152x2048"}],"required":false,"description":"生成图像尺寸","default_value":"2048x2048"},{"in":"body","name":"width","step":8,"type":"integer","control":"number","maximum":2048,"minimum":1024,"required":false,"description":"生成图片的宽度，最大为2048，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"height","step":8,"type":"integer","control":"number","maximum":2048,"minimum":1024,"required":false,"description":"生成图片的高度， 最大为2048，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"guidance_scale","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"调节模型在生成图像时对输入提示的遵从程度，数值越大，遵从越严格","default_value":7.5},{"in":"body","name":"seed","type":"integer","control":"number","required":false,"description":"用于确保随机过程的结果可复现，从而保证实验的一致性。","default_value":42},{"in":"body","name":"prompt","type":"string","control":"string","samples":[{"label":"机器人","value":"机器人坐在露天草地上，正在画布上画画"},{"label":"女孩","value":"一个可爱的日本女孩，正面全身，坐在汽车左侧乘客座位上，系着安全带，用索尼 α7R IV 和电影灯光拍摄，超现实主义摄影"},{"label":"卡通角色","value":"身穿和服的卡通武士插图，正面，全身，带花朵图案， 2D 对称构图"}],"category":"main","required":true},{"in":"body","name":"lora_weights","type":"array","control":"array","required":false,"max_length":4,"min_length":0,"description":"lora权重配置"},{"in":"body","name":"lora_weights.url","type":"string","control":"text","required":false,"description":"lora模型权重的url，若为空则视为基础sdxl模型的推理。"},{"in":"body","name":"lora_weights.weight","type":"number","control":"number","required":false,"description":"若存在多个lora，此处填写长度须与lora数量相同，否则报错。这些数值用于不同lora权重之间的加权融合。","default_value":1},{"in":"body","name":"lora_scale","type":"number","control":"number","required":false,"description":"融合的lora权重对基础模型的影响程度。"},{"in":"body","name":"negative_prompt","type":"string","control":"text","required":false,"description":"负向提示词"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1752484998000,"updated_at":1774611813000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[377],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":383,"account":0,"account_type":0,"type":"text2image","service_id":1512,"service_type":"serverless","model_id":17019,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"文生图异步(1024*1024)","description":"支持生成 1024*1024 内分辨率的图像，支持动态尺寸","path":"v1/async/images/generations","protocol":"http","price":0.0050,"average_price":0.0050,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":3,"params":[{"in":"body","name":"size","type":"string","control":"select","options":[{"label":"1:1 (256*256)","value":"256x256"},{"label":"1:1 (512*512)","value":"512x512"},{"label":"1:1 (1024*1024)","value":"1024x1024"},{"label":"4:3 (1024*768)","value":"1024x768"},{"label":"3:4 (768*1024)","value":"768x1024"},{"label":"16:9 (1024*576)","value":"1024x576"},{"label":"9:16 (576*1024)","value":"576x1024"},{"label":"3:2 (1024*640)","value":"1024x640"},{"label":"2:3(640*1024)","value":"640x1024"}],"required":false,"description":"图片尺寸","default_value":"1024x1024"},{"in":"body","name":"guidance_scale","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"调节模型在生成图像时对输入提示的遵从程度，数值越大，遵从越严格","default_value":7.5},{"in":"body","name":"seed","type":"integer","control":"number","required":false,"description":"用于确保随机过程的结果可复现，从而保证实验的一致性。","default_value":42},{"in":"body","name":"prompt","type":"string","control":"string","samples":[{"label":"机器人","value":"机器人坐在露天草地上，正在画布上画画"},{"label":"女孩","value":"一个可爱的日本女孩，正面全身，坐在汽车左侧乘客座位上，系着安全带，用索尼 α7R IV 和电影灯光拍摄，超现实主义摄影"},{"label":"卡通角色","value":"身穿和服的卡通武士插图，正面，全身，带花朵图案， 2D 对称构图"}],"category":"main","required":true},{"in":"body","name":"lora_weights","type":"array","control":"array","required":false,"max_length":4,"min_length":0,"description":"lora权重配置"},{"in":"body","name":"lora_weights.url","type":"string","control":"text","required":false,"description":"lora模型权重的url，若为空则视为基础sdxl模型的推理。"},{"in":"body","name":"lora_weights.weight","type":"number","control":"number","required":false,"description":"若存在多个lora，此处填写长度须与lora数量相同，否则报错。这些数值用于不同lora权重之间的加权融合。","default_value":1},{"in":"body","name":"lora_scale","type":"number","control":"number","required":false,"description":"融合的lora权重对基础模型的影响程度。"},{"in":"body","name":"num_images_per_prompt","type":"integer","control":"number","maximum":4,"minimum":1,"required":false,"description":"生成图片张数，计费将根据生成张数*单价计费。"},{"in":"body","name":"width","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的宽度，最大为1024，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"height","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的高度， 最大为1024，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"response_format","type":"string","control":"select","options":[{"label":"b64_json","value":"b64_json"},{"label":"url","value":"url"}],"required":false,"description":"生成图像时返回的格式。必须为 url 或 b64_json 之一。","default_value":"url"},{"in":"body","name":"negative_prompt","type":"string","control":"text","required":false,"description":"负向提示词"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1753252608000,"updated_at":1765192759000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[383],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":1227},{"id":384,"account":0,"account_type":0,"type":"text2image","service_id":1512,"service_type":"serverless","model_id":17019,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"文生图异步(1536*1536)","description":"支持生成 1536*1536 内分辨率的图像，支持动态尺寸","path":"v1/async/images/generations","protocol":"http","price":0.0100,"average_price":0.0100,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":4,"params":[{"in":"body","name":"size","type":"string","control":"select","options":[{"label":"1:1 (1536*1536)","value":"1536x1536"},{"label":"4:3 (1536*1152)","value":"1536x1152"},{"label":"3:4 (1152*1536)","value":"1152x1536"},{"label":"16:9 (1536*864)","value":"1536x864"},{"label":"9:16 (864*1536)","value":"864x1536"},{"label":"3:2 (1536*1024)","value":"1536x1024"},{"label":"2:3 (1024*1536)","value":"1024x1536"}],"required":false,"description":"图片尺寸","default_value":"1536x1536"},{"in":"body","name":"guidance_scale","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"调节模型在生成图像时对输入提示的遵从程度，数值越大，遵从越严格","default_value":7.5},{"in":"body","name":"seed","type":"integer","control":"number","required":false,"description":"用于确保随机过程的结果可复现，从而保证实验的一致性。","default_value":42},{"in":"body","name":"prompt","type":"string","control":"string","samples":[{"label":"机器人","value":"机器人坐在露天草地上，正在画布上画画"},{"label":"女孩","value":"一个可爱的日本女孩，正面全身，坐在汽车左侧乘客座位上，系着安全带，用索尼 α7R IV 和电影灯光拍摄，超现实主义摄影"},{"label":"卡通角色","value":"身穿和服的卡通武士插图，正面，全身，带花朵图案， 2D 对称构图"}],"category":"main","required":true},{"in":"body","name":"lora_weights","type":"array","control":"array","required":false,"max_length":4,"min_length":0,"description":"lora权重配置"},{"in":"body","name":"lora_weights.url","type":"string","control":"text","required":false,"description":"lora模型权重的url，若为空则视为基础sdxl模型的推理。"},{"in":"body","name":"lora_weights.weight","type":"number","control":"number","required":false,"description":"若存在多个lora，此处填写长度须与lora数量相同，否则报错。这些数值用于不同lora权重之间的加权融合。","default_value":1},{"in":"body","name":"lora_scale","type":"number","control":"number","required":false,"description":"融合的lora权重对基础模型的影响程度。"},{"in":"body","name":"num_images_per_prompt","type":"integer","control":"number","maximum":4,"minimum":1,"required":false,"description":"生成图片张数，计费将根据生成张数*单价计费。"},{"in":"body","name":"width","step":8,"type":"integer","control":"number","maximum":1536,"minimum":8,"required":false,"description":"生成图片的宽度，最大为1536，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"height","step":8,"type":"integer","control":"number","maximum":1536,"minimum":8,"required":false,"description":"生成图片的高度， 最大为1536，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"response_format","type":"string","control":"select","options":[{"label":"b64_json","value":"b64_json"},{"label":"url","value":"url"}],"required":false,"description":"生成图像时返回的格式。必须为 url 或 b64_json 之一。","default_value":"url"},{"in":"body","name":"negative_prompt","type":"string","control":"text","required":false,"description":"负向提示词"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1753252726000,"updated_at":1764919290000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[384],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":1227},{"id":443,"account":0,"account_type":0,"type":"text2image","service_id":1512,"service_type":"serverless","model_id":17019,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"文生图异步(2048*2048)","description":"支持生成 2048*2048 内分辨率的图像，支持动态尺寸","path":"v1/async/images/generations","protocol":"http","price":0.0120,"average_price":0.0120,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":5,"params":[{"in":"body","name":"size","type":"string","control":"select","options":[{"label":"1:1 (2048×2048)","value":"2048x2048"},{"label":"4:3 (2048×1536)","value":"2048x1536"},{"label":"3:4 (1536×2048)","value":"1536x2048"},{"label":"3:2 (2048×1360)","value":"2048x1360"},{"label":"2:3 (1360×2048)","value":"1360x2048"},{"label":"16:9 (2048×1152)","value":"2048x1152"},{"label":"9:16 (1152×2048)","value":"1152x2048"}],"required":false,"description":"生成图像尺寸","default_value":"2048x2048"},{"in":"body","name":"guidance_scale","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"调节模型在生成图像时对输入提示的遵从程度，数值越大，遵从越严格","default_value":7.5},{"in":"body","name":"seed","type":"integer","control":"number","required":false,"description":"用于确保随机过程的结果可复现，从而保证实验的一致性。","default_value":42},{"in":"body","name":"prompt","type":"string","control":"string","samples":[{"label":"机器人","value":"机器人坐在露天草地上，正在画布上画画"},{"label":"女孩","value":"一个可爱的日本女孩，正面全身，坐在汽车左侧乘客座位上，系着安全带，用索尼 α7R IV 和电影灯光拍摄，超现实主义摄影"},{"label":"卡通角色","value":"身穿和服的卡通武士插图，正面，全身，带花朵图案， 2D 对称构图"}],"category":"main","required":true},{"in":"body","name":"lora_weights","type":"array","control":"array","required":false,"max_length":4,"min_length":0,"description":"lora权重配置"},{"in":"body","name":"lora_weights.url","type":"string","control":"text","required":false,"description":"lora模型权重的url，若为空则视为基础sdxl模型的推理。"},{"in":"body","name":"lora_weights.weight","type":"number","control":"number","required":false,"description":"若存在多个lora，此处填写长度须与lora数量相同，否则报错。这些数值用于不同lora权重之间的加权融合。","default_value":1},{"in":"body","name":"lora_scale","type":"number","control":"number","required":false,"description":"融合的lora权重对基础模型的影响程度。"},{"in":"body","name":"num_images_per_prompt","type":"integer","control":"number","maximum":4,"minimum":1,"required":false,"description":"生成图片张数，计费将根据生成张数*单价计费。"},{"in":"body","name":"width","step":8,"type":"integer","control":"number","maximum":2048,"minimum":1024,"required":false,"description":"生成图片的宽度，最大为2048，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"height","step":8,"type":"integer","control":"number","maximum":2048,"minimum":1024,"required":false,"description":"生成图片的高度， 最大为2048，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"response_format","type":"string","control":"select","options":[{"label":"b64_json","value":"b64_json"},{"label":"url","value":"url"}],"required":false,"description":"生成图像时返回的格式。必须为 url 或 b64_json 之一。","default_value":"url"},{"in":"body","name":"negative_prompt","type":"string","control":"text","required":false,"description":"负向提示词"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1757660465000,"updated_at":1764919290000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[443],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":134,"account":0,"account_type":0,"type":"text2image","service_id":1512,"service_type":"serverless","model_id":17019,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"人脸迁移","description":"支持生成 1024*1024 内分辨率的图像，支持动态尺寸,将生成图片中的人物换脸","path":"v1/images/face-migration","protocol":"http","price":0.0050,"average_price":0.0050,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":18,"params":[{"in":"form","name":"image","type":"string","label":"图像","control":"file","category":"main","required":true,"description":"图像"},{"in":"form","name":"size","type":"string","control":"select","options":[{"label":"1:1 (256*256)","value":"256x256"},{"label":"1:1 (512*512)","value":"512x512"},{"label":"1:1 (1024*1024)","value":"1024x1024"},{"label":"4:3 (1024*768)","value":"1024x768"},{"label":"3:4 (768*1024)","value":"768x1024"},{"label":"16:9 (1024*576)","value":"1024x576"},{"label":"9:16 (576*1024)","value":"576x1024"},{"label":"3:2 (1024*683)","value":"1024x683"},{"label":"2:3(683*1024)","value":"683x1024"}],"required":false,"description":"图片尺寸","default_value":"1024x1024"},{"in":"form","name":"guidance_scale","step":0.1,"type":"number","control":"slider","maximum":10,"minimum":1,"required":false,"description":"调节模型在生成图像时对输入提示的遵从程度，数值越大，遵从越严格","default_value":4},{"in":"form","name":"num_inference_steps","step":1,"type":"number","control":"slider","maximum":8,"minimum":0,"required":false,"description":"生成图像时模型执行迭代的次数。数值越大，生成的图像质量会更高","default_value":4},{"in":"form","name":"id_weight","step":1,"type":"number","control":"slider","maximum":3,"minimum":0,"required":false,"description":"参考图像的人物面部对生成图片的影响程度。","default_value":1},{"in":"form","name":"id_timestep_start","step":1,"type":"number","control":"slider","maximum":8,"minimum":0,"required":false,"description":"从该时间步开始进行人脸嵌入，开始的时间步早会使得模型更早考虑面部特征。","default_value":0},{"in":"form","name":"seed","type":"integer","control":"number","required":false,"description":"用于确保随机过程的结果可复现，从而保证实验的一致性。"},{"in":"form","name":"width","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的宽度，最大为1024，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"form","name":"height","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的高度， 最大为1024，只支持8的倍数。与 size 参数互斥，二者只需传其一。"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1735647217000,"updated_at":1782713217000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[134],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":864,"ident":"glm-4-9b-chat","type":"serverless","vendor":"vnet","region":"prod-vn-bj-ip","compute":"vnet/prod-vn-bj-ip/serverless_api/gpu/2.xlarge/vnet","model":14354,"name":"GLM-4-9B-Chat","remark":"GLM-4-9B-Chat 在语义、数学、推理、代码和知识等多方面均表现出较高性能。还具备网页浏览、代码执行、自定义工具调用和长文本推理。 支持包括日语，韩语，德语在内的 26 种语言。","price":0.0000,"props":{"name":"glm-4-9b-chat","service_type":"serverless"},"inventories":0,"created_at":1722924390000,"updated_at":1766128344000,"sort_order":11,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-vs79e29q6-avatar-zhipuai.png","description_html":"","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":39,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1150,"name":"32K","slug":"context_size_32k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-f8crosrdk-id-y4z8cy4z9-context_size.png","en_name":"32K","color":"blue","description":"上下文大小","href":"","created_at":1734693847615,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":14354,"name":"glm-4-9b-chat","namespace":"hf-models","mirror_namespace":"THUDM","mirror_url":"https://huggingface.co/THUDM/glm-4-9b-chat"},"total_calls":61487,"operation_summary":{"service_id":864,"unit_tag":0,"min_price":0.0000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0000,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":2,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"vnet_prod-vn-bj-ip_vnet/prod-vn-bj-ip/serverless_api/gpu/2.xlarge/vnet","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":681,"account":0,"account_type":0,"type":"completions","service_id":864,"service_type":"serverless","model_id":14354,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"文本补全","description":"与 OpenAI 补全接口兼容，上下文大小为 32K，擅长代码补全任务","path":"v1/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":1,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":512},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"frequency_penalty","type":"number","control":"number","required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户。"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1775202024000,"updated_at":1787554220000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[681],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":680,"account":0,"account_type":0,"type":"text2text","service_id":864,"service_type":"serverless","model_id":14354,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"文本对话","description":"与 OpenAI 对话接口兼容，上下文大小为 32K，性能好","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":2,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":512},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","type":"number","control":"number","required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低","default_value":1},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1775201992000,"updated_at":1787554235000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[680],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":400,"ident":"Kolors","type":"serverless","vendor":"vnet","region":"prod-vn-bj-ip","compute":"vnet/prod-vn-bj-ip/serverless_api/gpu/2.xlarge/vnet","model":16591,"name":"Kolors","remark":"Kolors 是由快手 Kolors 团队开发的文生图模型。由数十亿的参数训练，在视觉质量、中文语义理解和文本渲染方面有显著优势。","price":0.0000,"props":{"name":"Kolors","service_type":"serverless"},"inventories":0,"created_at":1721177231000,"updated_at":1765264683000,"sort_order":12,"status":1,"recommend":false,"cover":"https://foruda.gitee.com/images/1725358697550204516/9bc2a8ac_3.png","description_html":"","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1148,"name":"图像生成与处理","slug":"image_generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-to-image.svg","en_name":"Image Generation","color":"","description":"","href":"","created_at":1734660944141,"sort_order":4,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1158,"name":"1024*1024","slug":"1024_1024_label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287708983,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1164,"name":"风格迁移","slug":"image_style_transfer_label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287927789,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1165,"name":"垫图","slug":"padding_image_label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287943717,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1166,"name":"局部重绘","slug":"inpainting_label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287955605,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1204,"name":"MindSpore","slug":"mindspore","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-6ey3e9tht-mindspore(1).png","en_name":"MindSpore","color":"blue","description":"支持使用 MindSpore 推理","href":"","created_at":1742544414967,"sort_order":0,"categories":[{"id":2,"parent_id":0,"name":"框架","slug":"library","created_at":1700107824458,"icon":"","sort_order":2,"en_name":""},{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":16591,"name":"Kolors","namespace":"hf-models","mirror_namespace":"Kwai-Kolors","mirror_url":"https://huggingface.co/Kwai-Kolors/Kolors"},"total_calls":36030,"operation_summary":{"service_id":400,"unit_tag":0,"min_price":0.0100,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0500,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":7},"available_services":[],"vendors":[],"vendor_region_compute":"vnet_prod-vn-bj-ip_vnet/prod-vn-bj-ip/serverless_api/gpu/2.xlarge/vnet","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":70,"account":0,"account_type":0,"type":"text2image","service_id":400,"service_type":"serverless","model_id":16591,"vendor":"iluvatar","region":"prod-il-sh-if","compute":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","name":"文生图(1024*1024)","description":"支持生成 1024*1024 内分辨率的图像，支持动态尺寸","path":"v1/images/generations","protocol":"http","price":0.0100,"average_price":0.0200,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"size","type":"string","control":"select","options":[{"label":"1:1 (256*256)","value":"256x256"},{"label":"1:1 (512*512)","value":"512x512"},{"label":"1:1 (1024*1024)","value":"1024x1024"},{"label":"4:3 (1024*768)","value":"1024x768"},{"label":"3:4 (768*1024)","value":"768x1024"},{"label":"16:9 (1024*576)","value":"1024x576"},{"label":"9:16 (576*1024)","value":"576x1024"},{"label":"3:2 (1024*640)","value":"1024x640"},{"label":"2:3(640*1024)","value":"640x1024"}],"required":false,"description":"图片尺寸","default_value":"1024x1024"},{"in":"body","name":"width","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的宽度，最大为1024，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"height","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的高度， 最大为1024，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"num_inference_steps","step":1,"type":"integer","control":"slider","maximum":30,"minimum":20,"required":false,"description":"生成图像时模型执行迭代的次数。数值越大，生成的图像质量会更高","default_value":25},{"in":"body","name":"guidance_scale","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"调节模型在生成图像时对输入提示的遵从程度，数值越大，遵从越严格","default_value":7.5},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1731311131000,"updated_at":1783078637000,"vendor_info":{"id":0,"name":"iluvatar","label":"天数智芯","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-lod1uv2gx-Group 62.svg"},"region_info":{"name":"prod-il-sh-if","label":"上海","props":{"products":["APP","ENDPOINT"]}},"compute_info":{"id":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","instance_size":"2.large","instance_type":"mr-v100","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":2,"cpus_num":12,"memory_gb":32,"architecture":"智铠 100","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-il-sh-if","vendor_name":"iluvatar","type":"SERVERLESS_API","storage":0},"ids":[70],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":66,"account":0,"account_type":0,"type":"text2image","service_id":0,"model_id":16591,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":66,"account":0,"account_type":0,"type":"text2image","service_id":400,"service_type":"serverless","model_id":16591,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"文生图(1024*1024)","description":"支持生成 1024*1024 内分辨率的图像，支持动态尺寸","path":"v1/images/generations","protocol":"http","price":0.0200,"average_price":0.0200,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":3,"params":[{"in":"body","name":"size","type":"string","control":"select","options":[{"label":"1:1 (256*256)","value":"256x256"},{"label":"1:1 (512*512)","value":"512x512"},{"label":"1:1 (1024*1024)","value":"1024x1024"},{"label":"4:3 (1024*768)","value":"1024x768"},{"label":"3:4 (768*1024)","value":"768x1024"},{"label":"16:9 (1024*576)","value":"1024x576"},{"label":"9:16 (576*1024)","value":"576x1024"},{"label":"3:2 (1024*683)","value":"1024x640"},{"label":"2:3(683*1024)","value":"640x1024"}],"required":false,"description":"图片尺寸","default_value":"1024x1024"},{"in":"body","name":"width","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的宽度，最大为1024，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"height","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的高度， 最大为1024，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"num_inference_steps","step":1,"type":"integer","control":"slider","maximum":30,"minimum":20,"required":false,"description":"生成图像时模型执行迭代的次数。数值越大，生成的图像质量会更高","default_value":25},{"in":"body","name":"guidance_scale","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"调节模型在生成图像时对输入提示的遵从程度，数值越大，遵从越严格","default_value":7.5},{"in":"body","name":"negative_prompt","type":"string","control":"text","required":false,"description":"指导模型避免生成所描述的内容。"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1731311131000,"updated_at":1782714487000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[66],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":70,"account":0,"account_type":0,"type":"text2image","service_id":0,"model_id":16591,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":68,"account":0,"account_type":0,"type":"image2image","service_id":400,"service_type":"serverless","model_id":16591,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"局部重绘","description":"支持自动识别并重绘图像中的部分区域，可用于移除或修复图像中的元素。","path":"v1/images/edits","protocol":"http","price":0.0500,"average_price":0.0500,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":4,"params":[{"in":"form","name":"image","type":"string","label":"图层","control":"file","category":"main","required":true,"description":"原图"},{"in":"form","name":"mask","type":"string","label":"图层","control":"file","category":"main","required":true,"description":"图层，用于指定图像的哪些部分需要进行修改或保持不变。"},{"in":"form","name":"prompt","type":"string","label":"提示词","control":"string","category":"main","required":true,"description":"需要生成图片的文案说明"},{"in":"form","name":"size","type":"string","control":"select","options":[{"label":"1:1 (256*256)","value":"256x256"},{"label":"1:1 (512*512)","value":"512x512"},{"label":"1:1 (1024*1024)","value":"1024x1024"},{"label":"4:3 (1024*768)","value":"1024x768"},{"label":"3:4 (768*1024)","value":"768x1024"},{"label":"16:9 (1024*576)","value":"1024x576"},{"label":"9:16 (576*1024)","value":"576x1024"},{"label":"3:2 (1024*640)","value":"1024x640"},{"label":"2:3(640*1024)","value":"640x1024"}],"required":false,"description":"生成文本语言","default_value":"1024x1024"},{"in":"form","name":"steps","step":1,"type":"integer","control":"slider","maximum":30,"minimum":20,"required":false,"description":"生成图像时模型执行迭代的次数。数值越大，生成的图像质量会更高","default_value":25},{"in":"form","name":"guidance_scale","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"调节模型在生成图像时对输入提示的遵从程度，数值越大，遵从越严格","default_value":6},{"in":"form","name":"negative_prompt","type":"string","control":"text","required":false,"description":"指导模型避免生成所描述的内容。"},{"in":"form","name":"seed","type":"string","control":"text","required":false,"description":"随机种子"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1731311131000,"updated_at":1782714512000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[68],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":69,"account":0,"account_type":0,"type":"image2image","service_id":0,"model_id":16591,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":69,"account":0,"account_type":0,"type":"image2image","service_id":400,"service_type":"serverless","model_id":16591,"vendor":"iluvatar","region":"prod-il-sh-if","compute":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","name":"局部重绘","description":"支持自动识别并重绘图像中的部分区域，可用于移除或修复图像中的元素。","path":"v1/images/edits","protocol":"http","price":0.0200,"average_price":0.0500,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":5,"params":[{"in":"form","name":"image","type":"string","label":"图层","control":"file","category":"main","required":true,"description":"原图"},{"in":"form","name":"mask","type":"string","label":"图层","control":"file","category":"main","required":true,"description":"图层，用于指定图像的哪些部分需要进行修改或保持不变。"},{"in":"form","name":"prompt","type":"string","label":"提示词","control":"string","category":"main","required":true,"description":"需要生成图片的文案说明"},{"in":"form","name":"size","type":"string","control":"select","options":[{"label":"1:1 (256*256)","value":"256x256"},{"label":"1:1 (512*512)","value":"512x512"},{"label":"1:1 (1024*1024)","value":"1024x1024"},{"label":"4:3 (1024*768)","value":"1024x768"},{"label":"3:4 (768*1024)","value":"768x1024"},{"label":"16:9 (1024*576)","value":"1024x576"},{"label":"9:16 (576*1024)","value":"576x1024"},{"label":"3:2 (1024*640)","value":"1024x640"},{"label":"2:3(640*1024)","value":"640x1024"}],"required":false,"description":"生成文本语言","default_value":"1024x1024"},{"in":"form","name":"steps","step":1,"type":"integer","control":"slider","maximum":30,"minimum":20,"required":false,"description":"生成图像时模型执行迭代的次数。数值越大，生成的图像质量会更高","default_value":25},{"in":"form","name":"guidance_scale","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"调节模型在生成图像时对输入提示的遵从程度，数值越大，遵从越严格","default_value":6},{"in":"form","name":"negative_prompt","type":"string","control":"text","required":false,"description":"指导模型避免生成所描述的内容。"},{"in":"form","name":"seed","type":"string","control":"text","required":false,"description":"随机种子"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1731311131000,"updated_at":1782714528000,"vendor_info":{"id":0,"name":"iluvatar","label":"天数智芯","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-lod1uv2gx-Group 62.svg"},"region_info":{"name":"prod-il-sh-if","label":"上海","props":{"products":["APP","ENDPOINT"]}},"compute_info":{"id":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","instance_size":"2.large","instance_type":"mr-v100","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":2,"cpus_num":12,"memory_gb":32,"architecture":"智铠 100","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-il-sh-if","vendor_name":"iluvatar","type":"SERVERLESS_API","storage":0},"ids":[69],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":68,"account":0,"account_type":0,"type":"image2image","service_id":0,"model_id":16591,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":392,"account":0,"account_type":0,"type":"image2image","service_id":400,"service_type":"serverless","model_id":16591,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"风格迁移","description":"支持将一张图像的风格快速应用到另一张图像，融合生成为一幅新图。","path":"v1/images/style-migration","protocol":"http","price":0.0200,"average_price":0.0200,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":6,"params":[{"in":"form","name":"prompt","type":"string","label":"提示词","control":"string","category":"main","required":true,"description":"需要生成图片的文案说明"},{"in":"form","name":"image","type":"file","accept":"image/*","control":"file","category":"main","required":true,"description":"指定需要应用的风格图像"},{"in":"form","name":"style_image","type":"file","label":"风格图","accept":"image/*","control":"file","category":"main","required":true,"description":"风格图"},{"in":"form","name":"size","type":"string","control":"select","options":[{"label":"1:1 (256*256)","value":"256x256"},{"label":"1:1 (512*512)","value":"512x512"},{"label":"1:1 (1024*1024)","value":"1024x1024"},{"label":"4:3 (1024*768)","value":"1024x768"},{"label":"3:4 (768*1024)","value":"768x1024"},{"label":"16:9 (1024*576)","value":"1024x576"},{"label":"9:16 (576*1024)","value":"576x1024"},{"label":"3:2 (1024*683)","value":"1024x683"},{"label":"2:3(683*1024)","value":"683x1024"}],"required":false,"description":"图片尺寸","default_value":"1024x1024"},{"in":"form","name":"steps","step":1,"type":"integer","control":"slider","maximum":30,"minimum":20,"required":false,"description":"生成图像时模型执行迭代的次数。数值越大，生成的图像质量会更高","default_value":25},{"in":"form","name":"guidance_scale","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"调节模型在生成图像时对输入提示的遵从程度，数值越大，遵从越严格","default_value":6},{"in":"form","name":"strength","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"参数控制生成图像对风格图的依赖程度，值越高风格化越强。","default_value":0.6},{"in":"form","name":"scale","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"参数控制风格效果的强弱，值越高风格化越明显。","default_value":0.5},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1753776747000,"updated_at":1782714545000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[392],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":205,"account":0,"account_type":0,"type":"image2image","service_id":0,"model_id":16591,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":393,"account":0,"account_type":0,"type":"text2image","service_id":400,"service_type":"serverless","model_id":16591,"vendor":"metax","region":"prod-mx-sh-ip","compute":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","name":"垫图","description":"支持根据输入的图片和文本描述，在原始图像的基础上进行调整、扩展或重新生成。","path":"v1/images/generations","protocol":"http","price":0.0200,"average_price":0.0200,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":11,"params":[{"in":"body","name":"image","type":"string","props":{"enableUrl":false},"accept":"image/*","control":"file","example":"参照图的 base64 字符串","category":"main","required":true,"description":"参照图，可根据您的要求生成相似的图片"},{"in":"body","name":"size","type":"string","control":"select","options":[{"label":"1:1 (256*256)","value":"256x256"},{"label":"1:1 (512*512)","value":"512x512"},{"label":"1:1 (1024*1024)","value":"1024x1024"},{"label":"4:3 (1024*768)","value":"1024x768"},{"label":"3:4 (768*1024)","value":"768x1024"},{"label":"16:9 (1024*576)","value":"1024x576"},{"label":"9:16 (576*1024)","value":"576x1024"},{"label":"3:2 (1024*683)","value":"1024x683"},{"label":"2:3(683*1024)","value":"683x1024"}],"required":false,"description":"图片尺寸","default_value":"1024x1024"},{"in":"body","name":"height","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的高度， 最大为1024，只支持8的倍数","default_value":1024},{"in":"body","name":"steps","step":1,"type":"integer","control":"slider","maximum":30,"minimum":20,"required":false,"description":"生成图像时模型执行迭代的次数。数值越大，生成的图像质量会更高","default_value":25},{"in":"body","name":"guidance_scale","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"调节模型在生成图像时对输入提示的遵从程度，数值越大，遵从越严格","default_value":6}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1753776940000,"updated_at":1782714564000,"vendor_info":{"id":0,"name":"metax","label":"沐曦","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-effotptlj-image 4.svg"},"region_info":{"name":"prod-mx-sh-ip","label":"上海","props":{"products":["SERVERLESS_API"]}},"compute_info":{"id":"metax/prod-mx-sh-ip/serverless_api/gpu/xlarge/metax","instance_size":"xlarge","instance_type":"metax","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":1,"cpus_num":128,"memory_gb":2048,"architecture":"曦云 C 系列","gpu_memory_gb":64,"region_name":"prod-mx-sh-ip","vendor_name":"metax","type":"SERVERLESS_API","storage":0},"ids":[393],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":205,"account":0,"account_type":0,"type":"image2image","service_id":400,"service_type":"serverless","model_id":16591,"vendor":"iluvatar","region":"prod-il-sh-if","compute":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","name":"风格迁移","description":"支持将一张图像的风格快速应用到另一张图像，融合生成为一幅新图。","path":"v1/images/style-migration","protocol":"http","price":0.0200,"average_price":0.0200,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":17,"params":[{"in":"form","name":"prompt","type":"string","label":"提示词","control":"string","category":"main","required":true,"description":"需要生成图片的文案说明"},{"in":"form","name":"image","type":"string","control":"file","category":"main","required":true,"description":"指定需要应用的风格图像"},{"in":"form","name":"style_image","type":"file","label":"风格图","accept":"image/*","control":"file","category":"main","required":true,"description":"风格图"},{"in":"form","name":"size","type":"string","control":"select","options":[{"label":"1:1 (256*256)","value":"256x256"},{"label":"1:1 (512*512)","value":"512x512"},{"label":"1:1 (1024*1024)","value":"1024x1024"},{"label":"4:3 (1024*768)","value":"1024x768"},{"label":"3:4 (768*1024)","value":"768x1024"},{"label":"16:9 (1024*576)","value":"1024x576"},{"label":"9:16 (576*1024)","value":"576x1024"},{"label":"3:2 (1024*683)","value":"1024x683"},{"label":"2:3(683*1024)","value":"683x1024"}],"required":false,"description":"图片尺寸","default_value":"1024x1024"},{"in":"form","name":"steps","step":1,"type":"integer","control":"slider","maximum":30,"minimum":20,"required":false,"description":"生成图像时模型执行迭代的次数。数值越大，生成的图像质量会更高","default_value":25},{"in":"form","name":"guidance_scale","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"调节模型在生成图像时对输入提示的遵从程度，数值越大，遵从越严格","default_value":6},{"in":"form","name":"strength","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"参数控制生成图像对风格图的依赖程度，值越高风格化越强。","default_value":0.6},{"in":"form","name":"scale","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"参数控制风格效果的强弱，值越高风格化越明显。","default_value":0.5},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1730324773000,"updated_at":1782714584000,"vendor_info":{"id":0,"name":"iluvatar","label":"天数智芯","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-lod1uv2gx-Group 62.svg"},"region_info":{"name":"prod-il-sh-if","label":"上海","props":{"products":["APP","ENDPOINT"]}},"compute_info":{"id":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","instance_size":"2.large","instance_type":"mr-v100","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":2,"cpus_num":12,"memory_gb":32,"architecture":"智铠 100","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-il-sh-if","vendor_name":"iluvatar","type":"SERVERLESS_API","storage":0},"ids":[205],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":392,"account":0,"account_type":0,"type":"image2image","service_id":0,"model_id":16591,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":134,"ident":"Qwen2-7B-Instruct","type":"serverless","vendor":"vnet","region":"prod-vn-bj-ip","compute":"vnet/prod-vn-bj-ip/serverless_api/gpu/2.xlarge/vnet","model":14490,"name":"Qwen2-7B-Instruct","remark":"Qwen2 是 Qwen 模型的最新系列，能够超越同等规模的最优开源模型甚至更大规模的模型，Qwen2 7B 在多个评测上取得显著的优势，尤其是代码及中文理解上。","price":0.0000,"props":{"name":"Qwen2-7B-Instruct","service_type":"serverless"},"inventories":0,"created_at":1719375020000,"updated_at":1750934891000,"sort_order":14,"status":1,"recommend":false,"cover":"https://gitee-ai.su.bcebos.com/v1/uploads/img/id-xegyngmvu-avatar-qwen.png","description_html":"","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":234,"name":"文本生成","slug":"text-generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-generation.svg","en_name":"Text Generation","color":"","description":"","href":"","created_at":1700443954730,"sort_order":36,"categories":[{"id":12,"parent_id":1,"name":"自然语言处理","slug":"nlp","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/nlp.svg","sort_order":12,"en_name":"Natural Language Processing"},{"id":18,"parent_id":8,"name":"自然语言处理","slug":"dataset-nlp","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-nlp.svg","sort_order":18,"en_name":"Natural Language Processing"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1153,"name":"24K","slug":"max_tokens_24k","icon":"https://gitee-ai.su.bcebos.com/v1/uploads/img/tag-icons/id-5v78q2mhf-context_size.png","en_name":"24K","color":"blue","description":"上下文大小","href":"","created_at":1734693869981,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":14490,"name":"Qwen2-7B-Instruct","namespace":"hf-models","mirror_namespace":"Qwen","mirror_url":"https://huggingface.co/Qwen/Qwen2-7B-Instruct"},"total_calls":22071,"operation_summary":{"service_id":134,"unit_tag":0,"min_price":0.0000,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0000,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":2,"operation_count":2},"available_services":[],"vendors":[],"vendor_region_compute":"vnet_prod-vn-bj-ip_vnet/prod-vn-bj-ip/serverless_api/gpu/2.xlarge/vnet","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":78,"account":0,"account_type":0,"type":"text2text","service_id":134,"service_type":"serverless","model_id":14490,"vendor":"iluvatar","region":"prod-il-sh-if","compute":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","name":"文本对话","description":"与 OpenAI 对话接口兼容，上下文大小为 24K，速度快","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":512},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","type":"number","control":"number","required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低"},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1732252945000,"updated_at":1775273954000,"vendor_info":{"id":0,"name":"iluvatar","label":"天数智芯","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-lod1uv2gx-Group 62.svg"},"region_info":{"name":"prod-il-sh-if","label":"上海","props":{"products":["APP","ENDPOINT"]}},"compute_info":{"id":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","instance_size":"2.large","instance_type":"mr-v100","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":2,"cpus_num":12,"memory_gb":32,"architecture":"智铠 100","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-il-sh-if","vendor_name":"iluvatar","type":"SERVERLESS_API","storage":0},"ids":[78],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":678,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":14490,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0},{"id":678,"account":0,"account_type":0,"type":"text2text","service_id":134,"service_type":"serverless","model_id":14490,"vendor":"enflame","region":"prod-enflame-qy-ip","compute":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","name":"文本对话","description":"与 OpenAI 对话接口兼容，上下文大小为 24K，速度快","path":"v1/chat/completions","protocol":"http","price":0.0000,"average_price":0.0000,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"stream","type":"boolean","control":"boolean","required":false,"description":"是否通过流式分批返回结果。如果设置为 true，生成过程中实时地向用户推送每一部分生成的文本。","default_value":true},{"in":"body","name":"max_tokens","type":"number","control":"number","required":false,"description":"模型可生成的最大 token 个数，不同模型上限不同。","default_value":512},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_p","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样方法的取值范围为 [0.0,1.0]。top_p 值确定模型从概率最高的前p%的候选词中选取 tokens；当 top_p 为 0 时，此参数无效。建议根据需求调整 top_p 或 temperature 参数，避免同时调整两者。","default_value":0.7},{"in":"body","name":"top_k","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"取值范围为 [0,100]，限制模型在每一步选择下一个词时，只从概率最高的前 k 个词中选取。数值越大，文本生成越多样。","default_value":50},{"in":"body","name":"frequency_penalty","type":"number","control":"number","required":false,"description":"用于调整自动生成文本中重复内容的频率。正数减少重复，负数增加重复。设置此参数后，如果一个词在文本中已经出现过，模型在后续生成中选择该词的概率会降低"},{"in":"body","name":"stop","type":"string","control":"text","required":false,"description":"模型在遇到 stop 所制定的字符时将停止生成。"},{"in":"body","name":"user","type":"string","control":"text","required":false,"description":"用于追踪和区分不同用户的对话请求。"},{"in":"head","name":"X-Failover-Enabled","type":"boolean","control":"boolean","required":false,"description":"是否启用故障转移机制：当前算力服务因故障而不可用时，允许切换至其他算力，以确保服务持续可用。","default_value":true}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1775187868000,"updated_at":1775187888000,"vendor_info":{"id":0,"name":"enflame","label":"燧原","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-dbfo2424x-cover-enflame.png"},"region_info":{"name":"prod-enflame-qy-ip","label":"甘肃庆阳","props":{}},"compute_info":{"id":"enflame/prod-enflame-qy-ip/serverless_api/gpu/8.xlarge/free","instance_size":"8.xlarge","instance_type":"free","price_per_hour":20.00,"status":"available","accelerator":"gpu","accelerators_num":8,"cpus_num":60,"memory_gb":448,"architecture":"S60","gpu_memory_gb":45,"region_name":"prod-enflame-qy-ip","vendor_name":"enflame","type":"SERVERLESS_API","storage":0},"ids":[678],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"failover_operations":[{"id":78,"account":0,"account_type":0,"type":"text2text","service_id":0,"model_id":14490,"price":0,"average_price":0,"input_million_tokens_price":0,"output_million_tokens_price":0,"input_million_tokens_price_average":0,"output_million_tokens_price_average":0,"billing_strategy":"standard","sort_order":0,"status":"available","ids":[],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"unit_tag":0}],"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":136,"ident":"stable-diffusion-3-medium","type":"serverless","vendor":"vnet","region":"prod-vn-bj-ip","compute":"vnet/prod-vn-bj-ip/serverless_api/gpu/2.xlarge/vnet","model":14628,"name":"stable-diffusion-3-medium","remark":"由 Stability AI 推出的最新文生图大模型。这一版本在继承了前代的优点上，对图像质量、文本理解和风格多样性等方面进行了显著改进，能够更准确地解读复杂的自然语言提示，并生成更为精确和多样化的图像。","price":0.0000,"props":{"name":"stable-diffusion-3-medium","service_type":"serverless"},"inventories":0,"created_at":1719403775000,"updated_at":1735288494000,"sort_order":16,"status":1,"recommend":false,"cover":"https://foruda.gitee.com/images/1725358782687324745/9331d279_3.png","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1148,"name":"图像生成与处理","slug":"image_generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-to-image.svg","en_name":"Image Generation","color":"","description":"","href":"","created_at":1734660944141,"sort_order":14,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1158,"name":"1024*1024","slug":"1024_1024_label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287708983,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":14628,"name":"stable-diffusion-3-medium","namespace":"hf-models","mirror_namespace":"stabilityai","mirror_url":"https://huggingface.co/stabilityai/stable-diffusion-3-medium"},"total_calls":6699,"operation_summary":{"service_id":136,"unit_tag":0,"min_price":0.0200,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0200,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"vnet_prod-vn-bj-ip_vnet/prod-vn-bj-ip/serverless_api/gpu/2.xlarge/vnet","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":102,"account":0,"account_type":0,"type":"text2image","service_id":136,"service_type":"serverless","model_id":14628,"vendor":"iluvatar","region":"prod-il-sh-if","compute":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","name":"文生图(1024*1024)","description":"支持生成 1024*1024 内分辨率的图像，支持动态尺寸","path":"v1/images/generations","protocol":"http","price":0.0200,"average_price":0.0200,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"size","type":"string","control":"select","options":[{"label":"1:1 (256*256)","value":"256x256"},{"label":"1:1 (512*512)","value":"512x512"},{"label":"1:1 (1024*1024)","value":"1024x1024"},{"label":"4:3 (1024*768)","value":"1024x768"},{"label":"3:4 (768*1024)","value":"768x1024"},{"label":"16:9 (1024*576)","value":"1024x576"},{"label":"9:16 (576*1024)","value":"576x1024"},{"label":"3:2 (1024*683)","value":"1024x683"},{"label":"2:3(683*1024)","value":"683x1024"}],"required":false,"description":"图片尺寸","default_value":"1024x1024"},{"in":"body","name":"num_inference_steps","step":1,"type":"integer","control":"slider","maximum":30,"minimum":20,"required":false,"description":"生成图像时模型执行迭代的次数。数值越大，生成的图像质量会更高","default_value":25},{"in":"body","name":"guidance_scale","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"调节模型在生成图像时对输入提示的遵从程度，数值越大，遵从越严格","default_value":7.5},{"in":"body","name":"negative_prompt","type":"string","control":"text","required":false,"description":"指导模型避免生成所描述的内容。"},{"in":"body","name":"width","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的宽度，最大为1024，只支持8的倍数。与 size 参数互斥，二者只需传其一。"},{"in":"body","name":"height","step":8,"type":"integer","control":"number","maximum":1024,"minimum":8,"required":false,"description":"生成图片的高度， 最大为1024，只支持8的倍数。与 size 参数互斥，二者只需传其一。"}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1733898202000,"updated_at":1782714619000,"vendor_info":{"id":0,"name":"iluvatar","label":"天数智芯","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-lod1uv2gx-Group 62.svg"},"region_info":{"name":"prod-il-sh-if","label":"上海","props":{"products":["APP","ENDPOINT"]}},"compute_info":{"id":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","instance_size":"2.large","instance_type":"mr-v100","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":2,"cpus_num":12,"memory_gb":32,"architecture":"智铠 100","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-il-sh-if","vendor_name":"iluvatar","type":"SERVERLESS_API","storage":0},"ids":[102],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":109,"ident":"stable-diffusion-xl-base-1.0","type":"serverless","vendor":"metax","region":"prod-mx-gz-ip","compute":"metax/prod-mx-gz-ip/serverless_api/gpu/2.xlarge/metax","model":11,"name":"stable-diffusion-xl-base-1.0","remark":"由 Stability AI 开发并开源的文生图大模型，其创意图像生成能力位居行业前列。具备出色的指令理解能力，能够支持反向 Prompt 定义来精确生成内容。","price":0.0000,"props":{"name":"stable-diffusion-xl-base-1.0","service_type":"serverless"},"inventories":0,"created_at":1718241079000,"updated_at":1761123187000,"sort_order":18,"status":1,"recommend":false,"cover":"https://foruda.gitee.com/images/1725358782687324745/9331d279_3.png","description_html":"","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":1148,"name":"图像生成与处理","slug":"image_generation","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-text-to-image.svg","en_name":"Image Generation","color":"","description":"","href":"","created_at":1734660944141,"sort_order":17,"categories":[{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1158,"name":"1024*1024","slug":"1024_1024_label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287708983,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]},{"id":1162,"name":"LoRA","slug":"lora","icon":"","en_name":"LoRA","color":"blue","description":"","href":"","created_at":1735287887794,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":11,"name":"stable-diffusion-xl-base-1.0","namespace":"hf-models","mirror_namespace":"stabilityai","mirror_url":"https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0"},"total_calls":10027,"operation_summary":{"service_id":109,"unit_tag":0,"min_price":0.0160,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0160,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"metax_prod-mx-gz-ip_metax/prod-mx-gz-ip/serverless_api/gpu/2.xlarge/metax","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":90,"account":0,"account_type":0,"type":"text2image","service_id":109,"service_type":"serverless","model_id":11,"vendor":"iluvatar","region":"prod-il-sh-if","compute":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","name":"文生图(1024*1024)","description":"支持生成 1024*1024 内分辨率的图像，支持动态尺寸,支持lora","path":"v1/images/generations","protocol":"http","price":0.0160,"average_price":0.0160,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"body","name":"size","type":"string","control":"select","options":[{"label":"1:1 (256*256)","value":"256x256"},{"label":"1:1 (512*512)","value":"512x512"},{"label":"1:1 (1024*1024)","value":"1024x1024"},{"label":"4:3 (1024*768)","value":"1024x768"},{"label":"3:4 (768*1024)","value":"768x1024"},{"label":"16:9 (1024*576)","value":"1024x576"},{"label":"9:16 (576*1024)","value":"576x1024"},{"label":"3:2 (1024*640)","value":"1024x640"},{"label":"2:3(640*1024)","value":"640x1024"}],"required":false,"description":"图片尺寸","default_value":"1024x1024"},{"in":"body","name":"num_inference_steps","step":1,"type":"number","control":"slider","maximum":50,"minimum":0,"required":false,"description":"生成图片的步数，其越高则图片精细度越高，步数增加存在边际效应。","default_value":25},{"in":"body","name":"guidance_scale","step":1,"type":"number","control":"slider","maximum":100,"minimum":0,"required":false,"description":"提示词引导系数，增大数值将引导生成图片的过程更加遵循提示词。","default_value":7.5},{"in":"body","name":"lora_weight_url","type":"string","control":"text","visible":false,"required":false,"description":"lora模型权重的url，若为空则视为基础sdxl模型的推理。"},{"in":"body","name":"lora_adapter_weights","type":"number","control":"number","visible":false,"required":false,"description":"若存在多个lora，此处填写长度须与lora数量相同，否则报错。这些数值用于不同lora权重之间的加权融合。"},{"in":"body","name":"lora_scale","type":"number","control":"number","required":false,"description":"融合的lora权重对基础模型的影响程度。"},{"in":"body","name":"seed","type":"integer","control":"number","required":false,"description":"用于确保随机过程的结果可复现，从而保证实验的一致性。"},{"in":"body","name":"negative_prompt","type":"string","control":"text","required":false,"description":"指导模型避免生成所描述的内容。"},{"in":"body","name":"fast_inference","type":"boolean","control":"boolean","required":false,"description":"若仅使用基础模型推理，开启该选项可大幅提高推理速度，同时保证图像质量。","default_value":true},{"in":"body","name":"lora_weights","type":"array","control":"array","required":false,"max_length":4,"min_length":0,"description":"lora权重配置"},{"in":"body","name":"lora_weights.url","type":"string","control":"text","required":false,"description":"lora模型权重的url，若为空则视为基础sdxl模型的推理。"},{"in":"body","name":"lora_weights.weight","type":"number","control":"number","required":false,"description":"若存在多个lora，此处填写长度须与lora数量相同，否则报错。这些数值用于不同lora权重之间的加权融合。","default_value":1}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1732847005000,"updated_at":1782714642000,"vendor_info":{"id":0,"name":"iluvatar","label":"天数智芯","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-lod1uv2gx-Group 62.svg"},"region_info":{"name":"prod-il-sh-if","label":"上海","props":{"products":["APP","ENDPOINT"]}},"compute_info":{"id":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","instance_size":"2.large","instance_type":"mr-v100","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":2,"cpus_num":12,"memory_gb":32,"architecture":"智铠 100","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-il-sh-if","vendor_name":"iluvatar","type":"SERVERLESS_API","storage":0},"ids":[90],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":111,"ident":"whisper-base","type":"serverless","vendor":"metax","region":"prod-mx-gz-ip","compute":"metax/prod-mx-gz-ip/serverless_api/gpu/2.xlarge/metax","model":446,"name":"whisper-base","remark":"由 Open AI 开发并开源自动语音识别（ASR）模型。能够很好地处理各种语言、口音和背景噪音。相较于 Large 版，Base 版处理速度更快，具有较少的参数。","price":0.0000,"props":{"name":"whisper-base","service_type":"serverless"},"inventories":0,"created_at":1718318862000,"updated_at":1786809584000,"sort_order":19,"status":1,"recommend":false,"cover":"https://foruda.gitee.com/images/1725358718989688790/c5d14244_3.png","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":240,"name":"自动语音识别","slug":"automatic-speech-recognition","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-automatic-speech-recognition.svg","en_name":"Automatic Speech Recognition","color":"","description":"","href":"","created_at":1700443954790,"sort_order":0,"categories":[{"id":13,"parent_id":1,"name":"音频","slug":"audio","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/audio.svg","sort_order":13,"en_name":"Audio"},{"id":19,"parent_id":8,"name":"音频","slug":"dataset-audio","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-audio.svg","sort_order":19,"en_name":"Audio"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1168,"name":"字幕","slug":"subtitles_label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287980705,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":446,"name":"whisper-base","namespace":"hf-models","mirror_namespace":"openai","mirror_url":"https://huggingface.co/openai/whisper-base"},"total_calls":7042,"operation_summary":{"service_id":111,"unit_tag":0,"min_price":0.0100,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0100,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"metax_prod-mx-gz-ip_metax/prod-mx-gz-ip/serverless_api/gpu/2.xlarge/metax","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":195,"account":0,"account_type":0,"type":"speech2text","service_id":111,"service_type":"serverless","model_id":446,"vendor":"iluvatar","region":"prod-il-sh-if","compute":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","name":"语音识别","description":"由 Open AI 开发并开源自动语音识别（ASR）模型。Whisper 由 68 万小时的多语言数据训练，其能够很好地处理各种语言、口音和背景噪音。","path":"v1/audio/transcriptions","protocol":"http","price":0.0100,"average_price":0.0100,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"form","name":"language","type":"string","control":"select","options":[{"label":"zh","value":"zh"},{"label":"en","value":"en"},{"label":"de","value":"de"},{"label":"es","value":"es"}],"required":false,"description":"生成文本语言","default_value":""},{"in":"form","name":"timestamp_granularities","type":"array","items":{"type":"string","control":"select","options":[{"label":"word","value":"word"},{"label":"segment","value":"segment"}]},"control":"multimodal","example":["word"],"category":"main","required":false,"max_length":2,"description":"用于控制转录结果的时间戳精度，可选择返回整句级或逐词级的时间标注。传segment时，response_format 必须是 verbose_json。传word时，response_format可以是 verbose_json、srt、vtt。两个值都传，response_format 必须是 verbose_json。"},{"in":"form","name":"response_format","type":"string","control":"select","options":[{"label":"verbose_json","value":"verbose_json"},{"label":"json","value":"json"},{"label":"text","value":"text"},{"label":"srt","value":"srt"},{"label":"vtt","value":"vtt"}],"required":false,"description":"响应格式","default_value":""},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。","default_value":0.7}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1734640856000,"updated_at":1775273954000,"vendor_info":{"id":0,"name":"iluvatar","label":"天数智芯","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-lod1uv2gx-Group 62.svg"},"region_info":{"name":"prod-il-sh-if","label":"上海","props":{"products":["APP","ENDPOINT"]}},"compute_info":{"id":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","instance_size":"2.large","instance_type":"mr-v100","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":2,"cpus_num":12,"memory_gb":32,"architecture":"智铠 100","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-il-sh-if","vendor_name":"iluvatar","type":"SERVERLESS_API","storage":0},"ids":[195],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]},{"service":{"id":110,"ident":"whisper-large","type":"serverless","vendor":"metax","region":"prod-mx-gz-ip","compute":"metax/prod-mx-gz-ip/serverless_api/gpu/2.xlarge/metax","model":150,"name":"whisper-large","remark":"由 Open AI 开发并开源自动语音识别（ASR）模型。Whisper 由 68 万小时的多语言数据训练，其能够很好地处理各种语言、口音和背景噪音。","price":0.0000,"props":{"name":"whisper-large","service_type":"serverless"},"inventories":0,"created_at":1718337228000,"updated_at":1786809584000,"sort_order":20,"status":1,"recommend":false,"cover":"https://foruda.gitee.com/images/1725358718989688790/c5d14244_3.png","visible":true,"purchasable":true,"access_scope":"PUBLIC","sale_mode":"one_time","tags":[{"id":240,"name":"自动语音识别","slug":"automatic-speech-recognition","icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/serverless-automatic-speech-recognition.svg","en_name":"Automatic Speech Recognition","color":"","description":"","href":"","created_at":1700443954790,"sort_order":1,"categories":[{"id":13,"parent_id":1,"name":"音频","slug":"audio","created_at":1700107824471,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/audio.svg","sort_order":13,"en_name":"Audio"},{"id":19,"parent_id":8,"name":"音频","slug":"dataset-audio","created_at":1700107824479,"icon":"https://gitee-ai.su.bcebos.com/v1/tag-icons/dataset-audio.svg","sort_order":19,"en_name":"Audio"},{"id":30,"parent_id":28,"name":"Serverless服务","slug":"serverless-service","created_at":1728529387723,"icon":"","sort_order":30,"en_name":""}]},{"id":1168,"name":"字幕","slug":"subtitles_label","icon":"","en_name":"","color":"blue","description":"","href":"","created_at":1735287980705,"sort_order":0,"categories":[{"id":31,"parent_id":28,"name":"Serverless服务标签","slug":"serverless-service-label","created_at":1728529387723,"icon":"","sort_order":31,"en_name":""}]}],"model_info":{"id":150,"name":"whisper-large","namespace":"hf-models","mirror_namespace":"openai","mirror_url":"https://huggingface.co/openai/whisper-large"},"total_calls":3528,"operation_summary":{"service_id":110,"unit_tag":0,"min_price":0.0150,"min_input_million_tokens_price":0E-8,"min_output_million_tokens_price":0E-8,"max_price":0.0150,"max_input_million_tokens_price":0E-8,"max_output_million_tokens_price":0E-8,"free_operation_count":0,"operation_count":1},"available_services":[],"vendors":[],"vendor_region_compute":"metax_prod-mx-gz-ip_metax/prod-mx-gz-ip/serverless_api/gpu/2.xlarge/metax","video_resolution_map":{},"all_vendors":false,"subscription_service":false},"operations":[{"id":194,"account":0,"account_type":0,"type":"speech2text","service_id":110,"service_type":"serverless","model_id":150,"vendor":"iluvatar","region":"prod-il-sh-if","compute":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","name":"语音识别","description":"支持多语言的语音识别","path":"v1/audio/transcriptions","protocol":"http","price":0.0150,"average_price":0.0150,"input_million_tokens_price":0E-8,"output_million_tokens_price":0E-8,"input_million_tokens_price_average":0E-8,"output_million_tokens_price_average":0E-8,"billing_strategy":"standard","sort_order":0,"params":[{"in":"form","name":"language","type":"string","control":"select","options":[{"label":"zh","value":"zh"},{"label":"en","value":"en"},{"label":"de","value":"de"},{"label":"es","value":"es"}],"required":false,"description":"生成文本语言","default_value":""},{"in":"form","name":"timestamp_granularities","type":"array","items":{"type":"string","control":"select","options":[{"label":"word","value":"word"},{"label":"segment","value":"segment"}]},"control":"multimodal","example":["word"],"category":"main","required":false,"max_length":2,"description":"用于控制转录结果的时间戳精度，可选择返回整句级或逐词级的时间标注。传segment时，response_format 必须是 verbose_json。传word时，response_format可以是 verbose_json、srt、vtt。两个值都传，response_format 必须是 verbose_json。"},{"in":"form","name":"response_format","type":"string","control":"select","options":[{"label":"verbose_json","value":"verbose_json"},{"label":"json","value":"json"},{"label":"text","value":"text"},{"label":"srt","value":"srt"},{"label":"vtt","value":"vtt"}],"required":false,"description":"响应格式","default_value":""},{"in":"body","name":"temperature","step":0.1,"type":"number","control":"slider","maximum":1,"minimum":0,"required":false,"description":"采样温度控制输出的随机性。温度值在 [0.0, 1.0] 范围内，值越高，输出越随机和创造性；值越低，输出越稳定。","default_value":0.7}],"status":"available","free_use":true,"api_format":"OPEN_AI","created_at":1734578881000,"updated_at":1775273954000,"vendor_info":{"id":0,"name":"iluvatar","label":"天数智芯","status":"available","ratio":0.70,"logo":"https://gitee-ai.su.bcebos.com/v1/uploads/img/vendor-icons/id-lod1uv2gx-Group 62.svg"},"region_info":{"name":"prod-il-sh-if","label":"上海","props":{"products":["APP","ENDPOINT"]}},"compute_info":{"id":"iluvatar/prod-il-sh-if/serverless_api/gpu/2.large/mr-v100","instance_size":"2.large","instance_type":"mr-v100","price_per_hour":0.00,"status":"available","accelerator":"gpu","accelerators_num":2,"cpus_num":12,"memory_gb":32,"architecture":"智铠 100","gpu_memory_gb":32,"gpu_performance_rate":"full","region_name":"prod-il-sh-if","vendor_name":"iluvatar","type":"SERVERLESS_API","storage":0},"ids":[194],"total":0,"success":0,"failed":0,"bytes":0,"took":0,"amount":0,"engines":["GiEngine"],"visible":true,"unit_tag":0}]}]