跳到内容

ChatGPT ​

更新时间

Endpoint: https://mmu.vod-qcloud.com/v1认证: VOD ApiToken / Bearer

ChatGPT 是文本生成 / 对话模型,通过兼容 OpenAI 的接口调用,支持 JSON 与 SSE 流式响应。支持 Chat Completions、Responses 与 Messages 协议,提供 GPT-4.x / GPT-5.x 多个系列,覆盖通用对话、代码与推理场景。

基本信息 ​

项目值
调用地址https://mmu.vod-qcloud.com
认证在 VOD 应用程序中签发的 ApiToken / Bearer
输出文本 / 结构化响应
响应方式JSON / SSE

调用模型 ​

模型 ID协议
gpt-chat-latestChat Completions / Responses / Messages
gpt-6.1-sol NEWChat Completions / Responses / Messages
gpt-6-solChat Completions / Responses / Messages
gpt-6-lunaChat Completions / Responses / Messages
gpt-6-astraChat Completions / Responses / Messages
gpt-5.6-terraChat Completions / Responses / Messages
gpt-5.6-solChat Completions / Responses / Messages
gpt-5.6-lunaChat Completions / Responses / Messages
gpt-5.5Chat Completions / Responses / Messages
gpt-5.4-proChat Completions / Responses / Messages
gpt-5.4-nanoChat Completions / Responses / Messages
gpt-5.4-miniChat Completions / Responses / Messages
gpt-5.4Chat Completions / Responses / Messages
gpt-5.3-codexChat Completions / Responses / Messages
gpt-5.3-chatChat Completions / Responses / Messages
gpt-5.2-chatChat Completions / Responses / Messages
gpt-5.2Chat Completions / Responses / Messages
gpt-5.1-chatChat Completions / Responses / Messages
gpt-5.1Chat Completions / Responses / Messages
gpt-5-nanoChat Completions / Responses / Messages
gpt-5-miniChat Completions / Responses / Messages
gpt-5-chatChat Completions / Responses / Messages
gpt-5Chat Completions / Responses / Messages
gpt-4oChat Completions / Responses / Messages
gpt-4.1Chat Completions / Responses / Messages

在 https://text-aigc.vod-qcloud.com/v1/chat/completions 中也可以用 VOD ApiToken 使用 JSON 响应与 SSE 流式。Chat Completions 流式用 stream=true 与 stream_options.include_usage=true 请求。

支持协议 ​

规范路径对象
Chat CompletionsPOST /v1/chat/completions下方所有模型
ResponsesPOST /v1/responses下方所有模型
MessagesPOST /v1/messages下方所有模型

本文档列出的所有 ChatGPT 模型都支持 WAND 网关的 Responses。各协议的输入字段遵循 API Reference。

各版本采样选项 ​

模型 IDtemperaturetop_p
gpt-chat-latest不支持不支持
gpt-6-solreasoning.effort=none 时可设置reasoning.effort=none 时可设置
gpt-6-lunareasoning.effort=none 时可设置reasoning.effort=none 时可设置
gpt-6-astra不支持不支持
gpt-5.6-terrareasoning.effort=none 时可设置reasoning.effort=none 时可设置
gpt-5.6-solreasoning.effort=none 时可设置reasoning.effort=none 时可设置
gpt-5.6-lunareasoning.effort=none 时可设置reasoning.effort=none 时可设置
gpt-5.5reasoning.effort=none 时可设置reasoning.effort=none 时可设置
gpt-5.4-pro不支持不支持
gpt-5.4-nano可设置可设置
gpt-5.4-mini可设置可设置
gpt-5.4可设置可设置
gpt-5.3-codex可设置可设置
gpt-5.3-chat不支持不支持
gpt-5.2-chat不支持不支持
gpt-5.2可设置可设置
gpt-5.1-chat不支持不支持
gpt-5.1可设置可设置
gpt-5-nano不支持不支持
gpt-5-mini不支持不支持
gpt-5-chat不支持不支持
gpt-5不支持不支持
gpt-4o可设置可设置
gpt-4.1可设置可设置

temperature 指定采样温度,top_p 指定累积概率范围。以一种方式为主进行调整。

各版本推理设置 ​

模型 ID设置方式禁用
gpt-chat-latest不支持推理设置—
gpt-6-solreasoning.effortnone
gpt-6-lunareasoning.effortnone
gpt-6-astrareasoning.effort无法禁用
gpt-5.6-terrareasoning.effortnone
gpt-5.6-solreasoning.effortnone
gpt-5.6-lunareasoning.effortnone
gpt-5.5reasoning.effortnone
gpt-5.4-proreasoning.effort无法禁用
gpt-5.4-nanoreasoning.effortnone
gpt-5.4-minireasoning.effortnone
gpt-5.4reasoning.effortnone
gpt-5.3-codexreasoning.effortnone
gpt-5.3-chat不支持推理设置—
gpt-5.2-chat不支持推理设置—
gpt-5.2reasoning.effortnone
gpt-5.1-chat不支持推理设置—
gpt-5.1reasoning.effortnone
gpt-5-nanoreasoning.effort无法禁用
gpt-5-minireasoning.effort无法禁用
gpt-5-chat不支持推理设置—
gpt-5reasoning.effort无法禁用
gpt-4o不支持推理设置—
gpt-4.1不支持推理设置—

Responses 的基准字段是 reasoning.effort。Chat Completions 中使用 reasoning_effort。无法禁用的模型请指定该版本允许的推理强度。

各版本推理及令牌上限 ​

模型 ID推理强度上下文令牌最大输出令牌
gpt-6-solnone / low / medium / high / xhigh / max1,050,000128,000
gpt-6-lunanone / low / medium / high / xhigh / max1,050,000128,000
gpt-6-astralow / medium / high / xhigh / max1,050,000128,000
gpt-5.6-terranone / low / medium / high / xhigh / max1,050,000128,000
gpt-5.6-solnone / low / medium / high / xhigh / max1,050,000128,000
gpt-5.6-lunanone / low / medium / high / xhigh / max1,050,000128,000
gpt-5.5none / low / medium / high / xhigh1,050,000128,000
gpt-5.4-promedium / high / xhigh1,050,000128,000
gpt-5.4-nanonone / low / medium / high / xhigh400,000128,000
gpt-5.4-mininone / low / medium / high / xhigh400,000128,000
gpt-5.4none / low / medium / high / xhigh1,050,000128,000
gpt-5.3-codex默认设置400,000128,000
gpt-5.2none / low / medium / high / xhigh400,000128,000
gpt-5.1none / low / medium / high400,000128,000
gpt-5-nano默认设置400,000128,000
gpt-5-mini默认设置400,000128,000
gpt-5minimal / low / medium / high400,000128,000
gpt-4o默认设置128,00016,384
gpt-4.1默认设置knowledge1,047,576

请求参数 ​

参数必填类型说明
model必填String调用模型及各版本选项表中的模型 ID
messages必填ArrayChat Completions / Messages 的对话输入
input必填String / ArrayResponses 的输入。代替 messages 使用
max_tokens可选IntegerChat Completions / Messages 的生成上限。在 Messages 中必填
max_output_tokens可选IntegerResponses 的生成上限
stream可选Booleantrue:SSE / false:JSON

请求示例 ​

GPT 6.1 Sol Chat CompletionsNEW ​

json
{
  "model": "gpt-6.1-sol",
  "messages": [
    {
      "role": "user",
      "content": "Compute 17 + 25. Reply with the number only."
    }
  ],
  "max_tokens": 256,
  "stream": false
}

响应 ​

json
{
  "object": "chat.completion",
  "model": "gpt-6.1-sol",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "42"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 19,
    "prompt_tokens_details": {
      "audio_tokens": 0,
      "cached_tokens": 0
    },
    "prompt_cached_tokens_details": {
      "audio_tokens": 0
    },
    "completion_tokens": 5,
    "completion_tokens_details": {
      "accepted_prediction_tokens": 0,
      "audio_tokens": 0,
      "reasoning_tokens": 0,
      "rejected_prediction_tokens": 0
    },
    "total_tokens": 24
  }
}

Chat Completions ​

json
{
  "model": "gpt-chat-latest",
  "messages": [
    {
      "role": "user",
      "content": "Reply with exactly OK and nothing else."
    }
  ],
  "max_tokens": 512,
  "stream": false
}

响应 ​

json
{
  "id": "01a0de31-615d-7014-9372-2ab6ab19e70e",
  "object": "chat.completion",
  "model": "gpt-chat-latest",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "OK"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 14,
    "prompt_tokens_details": {
      "audio_tokens": 0,
      "cached_tokens": 0
    },
    "prompt_cached_tokens_details": {
      "audio_tokens": 0
    },
    "completion_tokens": 17,
    "completion_tokens_details": {
      "accepted_prediction_tokens": 0,
      "audio_tokens": 0,
      "reasoning_tokens": 0,
      "rejected_prediction_tokens": 0
    },
    "total_tokens": 31
  }
}

Responses ​

json
{
  "model": "gpt-6-astra",
  "input": "Reply with exactly OK and nothing else.",
  "max_output_tokens": 2048,
  "stream": false,
  "reasoning": {
    "effort": "low"
  }
}

响应 ​

json
{
  "id": "resp_06ba8773e2ed38cf006ab7df36befc8195ab2906ff31076da8",
  "object": "response",
  "model": "gpt-6-astra",
  "status": "completed",
  "output": [
    {
      "id": "msg_06ba8773e2ed38cf006ab7df37be88819585df66bae9e722a1",
      "type": "message",
      "status": "completed",
      "content": [
        {
          "type": "output_text",
          "annotations": ,
          "logprobs": ,
          "text": "OK"
        }
      ],
      "phase": "final_answer",
      "role": "assistant"
    }
  ],
  "usage": {
    "input_tokens": 14,
    "input_tokens_details": {
      "cache_write_tokens": 0,
      "cached_tokens": 0
    },
    "output_tokens": 5,
    "output_tokens_details": {
      "reasoning_tokens": 0
    },
    "total_tokens": 19
  }
}

Messages ​

json
{
  "model": "gpt-chat-latest",
  "messages": [
    {
      "role": "user",
      "content": "Reply with exactly OK and nothing else."
    }
  ],
  "max_tokens": 512,
  "stream": false
}

响应 ​

json
{
  "id": "01a0de34-7483-76a8-a2e6-5f997dc8388c",
  "model": "gpt-chat-latest",
  "type": "message",
  "role": "assistant",
  "content": [
    {
      "text": "OK",
      "type": "text"
    }
  ],
  "stop_reason": "end_turn",
  "usage": {
    "input_tokens": 14,
    "output_tokens": 5
  }
}

特殊设置 ​

Chat Completions 使用 reasoning_effort,Responses 使用 reasoning.effort。推理设置请按模型版本指定。gpt-6-astra 与 Pro 模型不使用关闭推理的 none,而是使用该版本允许的强度。

Responses 的 output 中可同时包含推理项与消息项。文本从 type=message 的 content 中读取 type=output_text。Chat Completions 读取 choices.message.content。

GPT-6 Sol 的工具调用使用 Responses 规范。Chat Completions 的函数调用用 reasoning_effort=none 构成。原生工具与网关兼容字段是分开的。

各版本附加选项 ​

gpt-5.4-pro 与 gpt-5.3-codex 通过 Responses 路径调用。用 reasoning.effort 指定推理强度。

基于 VitePress 构建 · 部署于腾讯云 EdgeOne Pages