创建消息请求（Claude）

简介

Claude 原生的消息接口，适用于 Claude Code 等原生 Anthropic 客户端。该接口遵循 Anthropic 的 API 规范，提供完整的 Claude 模型功能支持，包括扩展思考（Extended Thinking）、工具调用等高级特性。如果您使用 OpenAI 兼容的客户端（如 OpenAI SDK），建议使用 /v1/chat/completions 接口。

认证

Bearer Token，如 Bearer sk-xxxxxxxxxx

请求参数

model

string

required

Claude 模型标识，如 claude-sonnet-4-5-20250929、claude-opus-4-5-20251101

messages

array

required

对话消息列表，每个元素包含 role（user/assistant）和 content

max_tokens

integer

required

最大生成 token 数，必须大于 0

system

string | array

系统提示词，支持字符串格式或数组格式（用于 Prompt Caching）

stream

boolean

是否启用流式输出

temperature

number

采样温度，范围 0-1

top_p

number

核采样参数，范围 0-1

top_k

integer

Top-k 采样参数

stop_sequences

array

自定义停止序列

thinking

object

扩展思考配置，包含 type 和 budget_tokens

tools

array

工具定义列表

基础示例

非流式请求
流式请求（SSE）
Python 示例

curl -X POST "https://console.mixroute.io/v1/messages" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-xxxxxxxxxx" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-sonnet-4-5-20250929",
    "max_tokens": 1024,
    "messages": [
      {"role": "user", "content": "请用中文简要介绍人工智能"}
    ]
  }'

curl -X POST "https://console.mixroute.io/v1/messages" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-xxxxxxxxxx" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-sonnet-4-5-20250929",
    "max_tokens": 1024,
    "stream": true,
    "messages": [
      {"role": "user", "content": "请用中文简要介绍人工智能"}
    ]
  }'

from anthropic import Anthropic

client = Anthropic(
    api_key="sk-xxxxxxxxxx",
    base_url="https://console.mixroute.io"
)

# 非流式
message = client.messages.create(
    model="claude-sonnet-4-5-20250929",
    max_tokens=1024,
    messages=[
        {"role": "user", "content": "请用中文简要介绍人工智能"}
    ]
)
print(message.content[0].text)

# 流式
with client.messages.stream(
    model="claude-sonnet-4-5-20250929",
    max_tokens=1024,
    messages=[
        {"role": "user", "content": "请用中文简要介绍人工智能"}
    ]
) as stream:
    for text_block in stream.text_stream:
        print(text_block, end="")

高级功能

系统提示词

字符串格式
数组格式

curl -X POST "https://console.mixroute.io/v1/messages" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-xxxxxxxxxx" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-sonnet-4-5-20250929",
    "max_tokens": 1024,
    "system": "你是一个专业的编程助手，擅长解释复杂的技术概念。",
    "messages": [
      {"role": "user", "content": "什么是递归？"}
    ]
  }'

curl -X POST "https://console.mixroute.io/v1/messages" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-xxxxxxxxxx" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-sonnet-4-5-20250929",
    "max_tokens": 1024,
    "system": [
      {
        "type": "text",
        "text": "你是一个专业的编程助手，擅长解释复杂的技术概念。"
      }
    ],
    "messages": [
      {"role": "user", "content": "什么是递归？"}
    ]
  }'

扩展思考（Extended Thinking）

基础用法
Python 示例

curl -X POST "https://console.mixroute.io/v1/messages" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-xxxxxxxxxx" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-sonnet-4-5-20250929",
    "max_tokens": 16000,
    "thinking": {
      "type": "enabled",
      "budget_tokens": 10000
    },
    "messages": [
      {"role": "user", "content": "给出一道中等难度的几何题并分步解析"}
    ]
  }'

from anthropic import Anthropic

client = Anthropic(
    api_key="sk-xxxxxxxxxx",
    base_url="https://console.mixroute.io"
)

with client.messages.stream(
    model="claude-sonnet-4-5-20250929",
    max_tokens=16000,
    thinking={
        "type": "enabled",
        "budget_tokens": 10000
    },
    messages=[
        {"role": "user", "content": "给出一道中等难度的几何题并分步解析"}
    ]
) as stream:
    for event in stream:
        if event.type == "content_block_delta":
            if hasattr(event.delta, "thinking"):
                print(f"[思考] {event.delta.thinking}", end="")
            elif hasattr(event.delta, "text"):
                print(event.delta.text, end="")

工具调用（Tools）

函数工具
Claude 官方网页搜索工具

curl -X POST "https://console.mixroute.io/v1/messages" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-xxxxxxxxxx" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-sonnet-4-5-20250929",
    "max_tokens": 1024,
    "tools": [
      {
        "name": "get_weather",
        "description": "根据城市获取天气信息",
        "input_schema": {
          "type": "object",
          "properties": {
            "city": {
              "type": "string",
              "description": "城市名称"
            }
          },
          "required": ["city"]
        }
      }
    ],
    "messages": [
      {"role": "user", "content": "上海的天气怎么样？"}
    ]
  }'

curl -X POST "https://console.mixroute.io/v1/messages" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-xxxxxxxxxx" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-sonnet-4-5-20250929",
    "max_tokens": 4096,
    "tools": [
      {
        "type": "web_search_20250305",
        "name": "web_search",
        "max_uses": 5
      }
    ],
    "messages": [
      {"role": "user", "content": "最近关于人工智能的新闻有哪些？"}
    ]
  }'

多模态输入（图像）

curl -X POST "https://console.mixroute.io/v1/messages" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-xxxxxxxxxx" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-sonnet-4-5-20250929",
    "max_tokens": 1024,
    "messages": [
      {
        "role": "user",
        "content": [
          {
            "type": "image",
            "source": {
              "type": "base64",
              "media_type": "image/jpeg",
              "data": "base64_encoded_image_data"
            }
          },
          {
            "type": "text",
            "text": "请描述这张图片"
          }
        ]
      }
    ]
  }'

Prompt Caching（提示词缓存）

通过缓存常用的上下文内容，可以显著降低成本和提升响应速度。缓存内容最小需要 1024 tokens。

System 缓存
Messages 缓存
Python SDK 示例

curl -X POST "https://console.mixroute.io/v1/messages" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-xxxxxxxxxx" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-sonnet-4-5-20250929",
    "max_tokens": 1024,
    "system": [
      {
        "type": "text",
        "text": "你是专业的文档分析助手。以下是需要分析的文档内容：[长文本内容，至少1024 tokens]",
        "cache_control": {"type": "ephemeral"}
      }
    ],
    "messages": [
      {"role": "user", "content": "请总结文档的主要观点"}
    ]
  }'

curl -X POST "https://console.mixroute.io/v1/messages" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-xxxxxxxxxx" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-sonnet-4-5-20250929",
    "max_tokens": 1024,
    "messages": [
      {
        "role": "user",
        "content": [
          {
            "type": "text",
            "text": "以下是需要分析的代码库：[大量代码内容，至少1024 tokens]",
            "cache_control": {"type": "ephemeral"}
          },
          {
            "type": "text",
            "text": "请找出代码中的潜在问题"
          }
        ]
      }
    ]
  }'

from anthropic import Anthropic

client = Anthropic(
    api_key="sk-xxxxxxxxxx",
    base_url="https://console.mixroute.io"
)

# System 缓存
message = client.messages.create(
    model="claude-sonnet-4-5-20250929",
    max_tokens=1024,
    system=[
        {
            "type": "text",
            "text": "你是专业的文档分析助手。以下是需要分析的文档内容：[长文本内容]",
            "cache_control": {"type": "ephemeral"}
        }
    ],
    messages=[
        {"role": "user", "content": "请总结文档的主要观点"}
    ]
)

# 查看缓存使用情况
print(f"缓存创建 tokens: {message.usage.cache_creation_input_tokens}")
print(f"缓存读取 tokens: {message.usage.cache_read_input_tokens}")

响应格式

非流式响应
流式响应

{
  "id": "msg_xxx",
  "type": "message",
  "role": "assistant",
  "content": [
    {
      "type": "text",
      "text": "回复内容..."
    }
  ],
  "model": "claude-sonnet-4-5-20250929",
  "stop_reason": "end_turn",
  "usage": {
    "input_tokens": 25,
    "output_tokens": 100,
    "cache_creation_input_tokens": 0,
    "cache_read_input_tokens": 0
  }
}

event: message_start
data: {"type":"message_start","message":{"id":"msg_xxx","type":"message","role":"assistant","content":[],"model":"claude-sonnet-4-5-20250929"}}

event: content_block_start
data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""}}

event: content_block_delta
data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"回复"}}

event: content_block_delta
data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"内容"}}

event: content_block_stop
data: {"type":"content_block_stop","index":0}

event: message_delta
data: {"type":"message_delta","delta":{"stop_reason":"end_turn"},"usage":{"output_tokens":100}}

event: message_stop
data: {"type":"message_stop"}

curl --request POST \
  --url https://console.mixroute.io/v1/messages \
  --header 'Authorization: Bearer sk-xxxxxxxxxx' \
  --header 'Content-Type: application/json' \
  --header 'anthropic-version: 2023-06-01' \
  --data '{
    "model": "claude-sonnet-4-5-20250929",
    "max_tokens": 1024,
    "messages": [
      {"role": "user", "content": "请用中文简要介绍人工智能"}
    ]
  }'

{
  "id": "msg_xxx",
  "type": "message",
  "role": "assistant",
  "content": [
    {
      "type": "text",
      "text": "人工智能是研究、开发用于模拟、延伸和扩展人的智能的理论、方法、技术及应用系统的一门新的技术科学..."
    }
  ],
  "model": "claude-sonnet-4-5-20250929",
  "stop_reason": "end_turn",
  "stop_sequence": null,
  "usage": {
    "input_tokens": 25,
    "output_tokens": 100
  }
}

当下热门模型

OpenAI 官方库使用

文本系列

文本向量系列

图像系列

视频系列

其他

创建消息请求（Claude）

简介

认证

请求参数

基础示例

高级功能

系统提示词

扩展思考（Extended Thinking）

工具调用（Tools）

多模态输入（图像）

Prompt Caching（提示词缓存）

响应格式

当下热门模型

OpenAI 官方库使用

文本系列

文本向量系列

图像系列

视频系列

其他

​简介

​认证

​请求参数

​基础示例

​高级功能

​系统提示词

​扩展思考（Extended Thinking）

​工具调用（Tools）

​多模态输入（图像）

​Prompt Caching（提示词缓存）

​响应格式

简介

认证

请求参数

基础示例

高级功能

系统提示词

扩展思考（Extended Thinking）

工具调用（Tools）

多模态输入（图像）

Prompt Caching（提示词缓存）

响应格式