对话补全

POST /v1/chat/completions

OpenAI 协议的主接口。绝大多数应用只需要这一个端点。

请求头

Authorization: Bearer ***
Content-Type: application/json

请求体

字段	类型	必填	说明
model	string	是	模型名，见 模型列表
messages	array	是	消息数组，见下
max_tokens	integer	否	输出上限
temperature	number	否	采样温度
top_p	number	否	核采样
stream	boolean	否	true 返回 SSE
stop	string/array	否	停止序列
tools	array	否	工具定义
tool_choice	string/object	否	工具选择策略
response_format	object	否	如 {"type":"json_object"}
user	string	否	终端用户标识（便于审计）

messages 结构

[
  {"role": "system", "content": "你是简洁的助手。"},
  {"role": "user", "content": "你好"},
  {"role": "assistant", "content": "你好，有什么可以帮你？"},
  {"role": "user", "content": "介绍一下你自己"}
]

多模态内容用数组：

{
  "role": "user",
  "content": [
    {"type": "text", "text": "这张图里是什么？"},
    {"type": "image_url", "image_url": {"url": "https://example.com/a.png"}}
  ]
}

最小请求

curl https://187api.com/v1/chat/completions \
  -H "Authorization: Bearer ***" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "MiniMax-M2.5",
    "messages": [{"role": "user", "content": "你好"}],
    "max_tokens": 128
  }'

非流式响应

{
  "id": "chatcmpl-xxx",
  "object": "chat.completion",
  "created": 1793532000,
  "model": "MiniMax-M2.5",
  "choices": [
    {
      "index": 0,
      "message": {"role": "assistant", "content": "你好！"},
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 12,
    "completion_tokens": 5,
    "total_tokens": 17
  }
}

:::info usage 里的缓存字段 启用缓存后，usage 里会出现缓存相关字段（如 prompt_tokens_details.cached_tokens）。 这是上游返回的真实用量，也是计费的依据。 :::

流式响应

请求体加 "stream": true：

curl -N https://187api.com/v1/chat/completions \
  -H "Authorization: Bearer ***" \
  -H "Content-Type: application/json" \
  -d '{"model":"MiniMax-M2.5","messages":[{"role":"user","content":"你好"}],"stream":true}'

data: {"choices":[{"delta":{"role":"assistant"},"index":0}]}

data: {"choices":[{"delta":{"content":"你"},"index":0}]}

data: {"choices":[{"delta":{"content":"好"},"index":0}]}

data: {"choices":[{"delta":{},"finish_reason":"stop","index":0}]}

data: [DONE]

工具调用

{
  "model": "MiniMax-M2.5",
  "messages": [{"role": "user", "content": "北京今天天气？"}],
  "tools": [
    {
      "type": "function",
      "function": {
        "name": "get_weather",
        "description": "查询城市天气",
        "parameters": {
          "type": "object",
          "properties": {"city": {"type": "string"}},
          "required": ["city"]
        }
      }
    }
  ],
  "tool_choice": "auto"
}

模型返回 finish_reason: "tool_calls" 时，把 tool_calls 结果按 {"role":"tool","tool_call_id":"...","content":"..."} 追加回 messages 再请求一次。

常见错误

错误	原因
model not found	分组里没有该模型，或模型名拼错
insufficient quota	令牌额度或余额不足
context length exceeded	输入超过模型上下文窗口
rate limit	触发限流