万物平台模型导入样例大全（实时更新中～）

引言

万物平台作为可支持多种 AI 模型导入的统一平台，其模型导入功能对开发者而言意义重大。本文将为大家详细介绍，该平台支持的一系列供应商，包括 OpenAI-API-compatible、联通元景、Ollama、通义千问、火山引擎等。同时，平台还支持 LLM、rerank、embedding 类型的模型导入。​
为了方便各位开发者判断所导入的模型是否被平台支持，本文会将所有的 http 请求一一列举出来，供大家参考使用。

LLM

OpenAI-API- compatible

curl --location 'URL' \
--header 'Content-Type: application/json' \
--header 'Authorization: Bearer api-key' \
--data '{"messages": [{"role": "user","content": "你是谁"}],"model": "gpt-oss-20b","stream": false
}'

非流式响应

{"id": "chatcmpl-ec8ed5bef88849c4ab3147f650b3d3ec","object": "chat.completion","created": 1754964185,"model": "gpt-oss-20b","choices": [{"index": 0,"message": {"role": "assistant","content": "您好！  \n我是 ChatGPT，OpenAI 研发的人工智能语言模型，能够根据输入的文本进行对话、回答问题、提供建议、撰写文章等。您有什么想聊或者想了解的，尽管告诉我吧！","tool_calls": []},"finish_reason": "stop"}],"usage": {"prompt_tokens": 73,"completion_tokens": 179,"total_tokens": 252}
}

流式响应

{"id": "endpoint_common_250891","object": "chat.completion.chunk","created": 1754964492,"model": "deepseek-r1","choices": [{"index": 0,"delta": {"role": "assistant","content": "吗"},"finish_reason": null}]
}

联通元景

curl --location 'https://maas.ai-yuanjing.com/openapi/compatible-mode/v1/chat/completions' \
--header 'Content-Type: application/json' \
--header 'Authorization: Bearer api-key' \
--data '{"messages": [{"role": "user","content": "你是谁"}],"model": "deepseek-r1","stream": false
}'

非流式响应

{"id": "endpoint_common_249820","object": "chat.completion","created": 1754965561,"model": "deepseek-r1","choices": [{"index": 0,"message": {"role": "assistant","content": "<think>\n嗯，用户问了一个很基础但也很关键的问题：“你是谁”。这可能是第一次使用智能助手的用户，也可能是想确认当前对话对象身份的老用户。\n\n用户可能带着试探或好奇的心态提问，背后或许隐藏着“你能帮我做什么”“我该怎么和你交流”这样的潜在需求。考虑到问题非常简短，用户可能处于两种状态：要么是匆忙测试功能，要么是带着观察态度谨慎接触新事物。\n\n这种情况下，回复需要同时做到三点：明确身份建立信任感，展示功能激发兴趣，用表情符号传递友好态度。要避免过于技术化的术语，比如不说“我是基于Transformer架构的LLM”，而是用“智能助手”这样通俗的表达。\n\n用户没有提供任何背景信息，所以采用中性称呼“你”最安全。最后那句“现在有什么可以帮你的吗”是重要的话术转折，既开放又带引导性，能自然推动对话进入实用阶段。\n</think>\n你好呀！我是 **DeepSeek-R1**，一款由中国的「深度求索（DeepSeek）」公司研发的智能助手。你可以把我当作一个知识丰富、乐于助人的聊天伙伴 😊\n\n我的本领包括：\n\n- 📚 回答各种知识性问题（历史、科技、文学、生活常识等）  \n- ✍️ 帮你写文章、改简历、起标题、写文案  \n- 📊 处理和分析文档（Word、PDF、Excel等）  \n- 🧠 提供学习辅导、解题思路、论文润色  \n- 📅 安排计划、整理资料、做读书笔记  \n- 💬 陪你聊天谈心、解闷放松  \n\n我目前是**完全免费的**，也没有语音功能（纯文字交流），但我支持超长上下文记忆（128K tokens），上传文档也没问题哦！\n\n现在有什么可以帮你的吗？🌟","tool_calls": null},"finish_reason": "stop"}],"usage": {"prompt_tokens": 6,"completion_tokens": 378,"total_tokens": 384}
}

流式响应

{"id": "endpoint_common_250934","object": "chat.completion.chunk","created": 1754965671,"model": "deepseek-r1","choices": [{"index": 0,"delta": {"role": "assistant","content": "🌟"},"finish_reason": null}]
}

Ollam

curl --location 'http://ip:11434/v1/chat/completions' \
--header 'Content-Type: application/json' \
--data '{"model": "deepseek-r1:1.5b","messages": [{"role": "user","content": "你是谁"}]
}'

非流式响应

{"id": "chatcmpl-403","object": "chat.completion","created": 1754965819,"model": "deepseek-r1:1.5b","system_fingerprint": "fp_ollama","choices": [{"index": 0,"message": {"role": "assistant","content": "<think>\n\n</think>\n\n您好！我是由中国的深度求索（DeepSeek）公司开发的智能助手DeepSeek-R1。如您有任何任何问题，我会尽我所能为您提供帮助。"},"finish_reason": "stop"}],"usage": {"prompt_tokens": 5,"completion_tokens": 40,"total_tokens": 45}
}

流式响应

{"id": "chatcmpl-323","object": "chat.completion.chunk","created": 1754965959,"model": "deepseek-r1:1.5b","system_fingerprint": "fp_ollama","choices": [{"index": 0,"delta": {"role": "assistant","content": "。"},"finish_reason": null}]
}

通义千问

curl --location 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions' \
--header 'Authorization: Bearer api-key' \
--header 'Content-Type: application/json' \
--data '{"model": "qwen-flash","messages": [{"role": "user","content": "你好"}],"stream": false,"enable_thinking":true 
}'

非流式响应

{"choices": [{"message": {"content": "你好！😊 有什么我可以帮你的吗？","reasoning_content": "嗯，用户发来“你好”，这是一个常见的打招呼方式。我需要友好地回应。首先，确认用户可能需要帮助，所以应该用中文回复，保持礼貌。\n\n接下来，考虑用户可能的意图。他们可能刚接触这个平台，或者有具体的问题需要解决。作为AI，我应该提供帮助，同时保持简洁。\n\n检查之前的对话历史，但这里没有上下文，所以需要从头开始。确保回复不带任何假设，避免复杂信息。\n\n可能需要询问用户需要什么帮助，这样能引导他们进一步说明需求。例如，“你好！有什么我可以帮你的吗？”\n\n还要注意语气，用表情符号可能让回复更亲切，但不确定用户偏好，所以可能保持中性。不过根据之前的例子，有时候用😊可以增加友好感。\n\n避免使用专业术语，保持口语化。比如不要说“请问您需要什么帮助？”，而是更自然的“你好！有什么我可以帮你的吗？”\n\n检查有没有拼写错误，确保回复正确。中文的“你好”是正确的。\n\n可能需要考虑用户是否是第一次使用，所以回复要友好且开放，鼓励他们提出问题。\n\n最后，确保回复简短，不要冗长，这样用户不会感到压力。\n\n所以，综合起来，回复应该是：“你好！😊 有什么我可以帮你的吗？” 这样既友好又明确。","role": "assistant"},"finish_reason": "stop","index": 0,"logprobs": null}],"object": "chat.completion","usage": {"prompt_tokens": 9,"completion_tokens": 292,"total_tokens": 301},"created": 1754966125,"system_fingerprint": null,"model": "qwen-flash","id": "chatcmpl-ae34fd13-c4c1-987d-bc20-09fe6f6d4936"
}

流式响应

{"choices": [{"delta": {"content": " 有什么我可以帮助","reasoning_content": null},"finish_reason": null,"index": 0,"logprobs": null}],"object": "chat.completion.chunk","usage": null,"created": 1754966192,"system_fingerprint": null,"model": "qwen-flash","id": "chatcmpl-7143c7c5-2128-9a6a-b8dc-18c2c306bc14"
}

火山引擎

curl --location 'https://ark.cn-beijing.volces.com/api/v3/chat/completions' \
--header 'Content-Type: application/json' \
--header 'Authorization: Bearer api-key' \
--data '{"do_sample": false,"messages": [{"content": "你好","role": "user"}],"model": "doubao-seed-1-6-flash-250715","repetition_penalty": 1.1,"stream": true,"temperature": 0.01
}'

非流式响应

{"choices": [{"finish_reason": "stop","index": 0,"logprobs": null,"message": {"content": "你好呀！很高兴见到你~ 有什么我可以帮你的吗？😊","reasoning_content": "\n用户现在说“你好”，我需要友好地回应。首先，保持礼貌和热情，直接回应用户的问候，然后可以稍微引导一下，问问用户有什么需要帮助的，这样能让对话继续下去。比如可以说“你好呀！很高兴见到你~ 有什么我可以帮你的吗？😊” 这样既回应了问候，又表达了愿意提供帮助的态度，还带了个表情符号显得更亲切。","role": "assistant"}}],"created": 1754966401,"id": "021754966399956ef34c300eb7cd27671b96789a585234f4250fd","model": "doubao-seed-1-6-flash-250715","service_tier": "default","object": "chat.completion","usage": {"completion_tokens": 119,"prompt_tokens": 85,"total_tokens": 204,"prompt_tokens_details": {"cached_tokens": 0},"completion_tokens_details": {"reasoning_tokens": 101}}
}

流式响应

{"choices": [{"delta": {"content": "😊","role": "assistant"},"index": 0}],"created": 1754966363,"id": "021754966363144ef34c300eb7cd27671b96789a585234fa38263","model": "doubao-seed-1-6-flash-250715","service_tier": "default","object": "chat.completion.chunk","usage": null
}

Rerank

OpenAI-API- compatible

curl --location 'URL' \
--header 'Authorization: Bearer api-key' \
--header 'Content-Type: application/json' \
--data '{"model": "模型id","query": "Apple","documents": ["你好啊，世界","我爱吃汉堡","尖尖我噶奶","乌萨奇","雷欧奥特曼"],"top_n": 2,"return_documents": true
}'

响应

{"model": "jina-reranker-v2-base-multilingual","usage": {"total_tokens": 42},"results": [{"index": 3,"document": {"text": "乌萨奇"},"relevance_score": 0.9674102663993835},{"index": 0,"document": {"text": "你好啊，世界"},"relevance_score": 0.16132023930549622},{"index": 1,"document": {"text": "我爱吃汉堡"},"relevance_score": 0.06853749603033066}]
}

联通元景

curl --location 'https://maas-api.ai-yuanjing.com/openapi/bge/v1/rerank' \
--header 'Content-Type: application/json' \
--header 'Authorization: Bearer api-key' \
--data '{"query": "今天天气不错","texts": ["天气很好","股票涨跌跟天气无关","今天北京雾霾严重"]
}'

响应

[{"index": 0,"score": 4.88899040222168,"document": "天气很好"},{"index": 1,"score": -4.119245529174805,"document": "股票涨跌跟天气无关"},{"index": 2,"score": -6.025328159332275,"document": "今天北京雾霾严重"}
]

通义千问

curl --location 'https://dashscope.aliyuncs.com/api/v1/services/rerank/text-rerank/text-rerank' \
--header 'Authorization: Bearer api-key' \
--header 'Content-Type: application/json' \
--data '{"model": "gte-rerank","input":{"query": "什么是文本排序模型","documents": ["文本排序模型广泛用于搜索引擎和推荐系统中，它们根据文本相关性对候选文本进行排序","量子计算是计算科学的一个前沿领域","预训练语言模型的发展给文本排序模型带来了新的进展"]},"parameters": {"return_documents": true,"top_n": 3}
}'

响应

{"output": {"results": [{"document": {"text": "文本排序模型广泛用于搜索引擎和推荐系统中，它们根据文本相关性对候选文本进行排序"},"index": 0,"relevance_score": 0.7315062086803763},{"document": {"text": "预训练语言模型的发展给文本排序模型带来了新的进展"},"index": 2,"relevance_score": 0.5831720487049298},{"document": {"text": "量子计算是计算科学的一个前沿领域"},"index": 1,"relevance_score": 0.04973238644524712}]},"usage": {"total_tokens": 79},"request_id": "8baa31dc-a786-9326-8e20-895c74bbff16"
}

Embedding

OpenAI-API- compatible

curl --location 'URL' \
--header 'Authorization: Bearer api-key' \
--header 'Content-Type: application/json' \
--data '{"model": "BAAI/bge-large-zh-v1.5","input": ["风急天高猿啸哀","渚清沙白鸟飞回"],"encoding_format": "float"
}'

响应

{"object": "list","data": [{"embedding": [0.018213777,-0.026578045,0.018492958,-0.01139059,0.04650041,0.021854298,-0.056059573,-0.034439784,0.043842606,……],"index": 1,"object": "embedding"}],"model": "BAAI/bge-large-zh-v1.5","usage": {"prompt_tokens": 18,"completion_tokens": 0,"total_tokens": 18}
}

联通元景

curl --location 'https://maas-api.ai-yuanjing.com/openapi/compatible-mode/v1/embeddings' \
--header 'Content-Type: application/json' \
--header 'Authorization: Bearer api-key' \
--data '{"model": "qwen3-embed-0.6b","input": [" 你好"],"encoding_format": "float"
}'

响应

{"id": "embd-15c7e7b5-eccf-9140-afe2-42e25e88fd6d","object": "list","created": 1754967486,"model": "qwen3-embed-0.6b","data": [{"object": "embedding","embedding": [0.013916015625,0.0147705078125,0.06689453125,……],"index": 0}],"usage": {"prompt_tokens": 3,"completion_tokens": 0,"total_tokens": 3}
}

Ollama

curl --location 'http://ip:11434/v1/embeddings' \
--header 'Content-Type: application/json' \
--data '{"encoding_format": "float","input": [" 你好"],"model": "granite-embedding:latest"
}'

响应

{"object": "list","data": [{"object": "embedding","embedding": [0.011519907,-0.14500324,0.092740096,……],"index": 0}],"model": "granite-embedding:latest","usage": {"prompt_tokens": 5,"total_tokens": 5}
}

通义千问

curl --location 'https://dashscope.aliyuncs.com/compatible-mode/v1/embeddings' \
--header 'Authorization: Bearer api-key' \
--header 'Content-Type: application/json' \
--data '{"encoding_format": "float","input": ["你好"],"model": "text-embedding-v4"}'

响应

{"data": [{"embedding": [0.00904093962162733,0.040408216416835785,-0.025859294459223747,0.03675258532166481,0.04067809507250786,……],"index": 0,"object": "embedding"}],"object": "list","model": "text-embedding-v4","usage": {"prompt_tokens": 2,"total_tokens": 2},"id": "21fe5ad3-c5b7-9ba4-bad2-9d926691b4f3"
}

OCR

联通元景

curl --location 'https://maas-api.ai-yuanjing.com/openapi/v1/unicom-ocr' \
--header 'Content-Type: multipart/form-data' \
--header 'Authorization: Bearer api-key' \
--form 'file=上传文件'

响应

{"id": "embd-15c7e7b5-eccf-9140-afe2-42e25e88fd6d","object": "list","created": 1754967486,"model": "qwen3-embed-0.6b","data": [{"object": "embedding","embedding": [0.013916015625,0.0147705078125,0.06689453125,……],"index": 0}],"usage": {"prompt_tokens": 3,"completion_tokens": 0,"total_tokens": 3}
}