Completions API

传统文本补全接口。接口根据 prompt 生成一个或多个文本候选，适用于兼容旧版 Completion 模型的场景。以下字段与 CompletionRequest 保持一致；网关不支持的模型或取值会返回 400。

该接口主要用于兼容旧版 Completion 模型。文中的 gpt-3.5-turbo-instruct 仅作为传统模型示例，实际可用模型以当前账号配置和上游渠道为准；新项目应优先评估 Chat Completions 或 Responses 接口。

请求参数

请求头

请求体

参数约束

• model 和 prompt 必须同时提供；prompt 为空数组会返回 400。
• temperature 与 top_p 均用于采样控制，通常只调整其中一个。
• stream=true 时，响应为 SSE 增量事件；stream=false（默认）时返回完整响应。
• n 应为正整数；max_tokens 不能为负数。实际上限由模型和网关配置决定。

请求体示例

{
  "model": "gpt-3.5-turbo-instruct",
  "prompt": "Write a short greeting:",
  "max_tokens": 64,
  "temperature": 0.7,
  "top_p": 1,
  "n": 1,
  "stop": ["\n\n"],
  "stream": false,
  "suffix": "",
  "echo": false
}

请求示例代码

curl -X POST "https://10000router.com/v1/completions" \
  -H "Authorization: Bearer $API_KEY" \
  -d '{
    "model": "gpt-3.5-turbo-instruct",
    "prompt": "Write a short greeting:",
    "max_tokens": 64,
    "temperature": 0.7,
    "stream": false
  }'

const response = await fetch("https://10000router.com/v1/completions", {
  method: "POST",
  headers: {
    Authorization: "Bearer " + process.env.API_KEY,
  },
  body: JSON.stringify({
    model: "gpt-3.5-turbo-instruct",
    prompt: "Write a short greeting:",
    max_tokens: 64,
    temperature: 0.7,
    stream: false
  })
});
console.log(await response.json());

package main

import (
  "log"
  "net/http"
  "os"
  "strings"
)

func main() {
  payload := `{"model":"gpt-3.5-turbo-instruct","prompt":"Write a short greeting:","max_tokens":64,"temperature":0.7,"stream":false}`
  req, err := http.NewRequest("POST", "https://10000router.com/v1/completions", strings.NewReader(payload))
  if err != nil { log.Fatal(err) }
  req.Header.Set("Authorization", "Bearer "+os.Getenv("API_KEY"))
  res, err := http.DefaultClient.Do(req)
  if err != nil { log.Fatal(err) }
  defer res.Body.Close()
}

import os
from openai import OpenAI

client = OpenAI(api_key=os.environ["API_KEY"], base_url="https://10000router.com/v1")

response = client.post(
    "https://10000router.com/v1/completions",
    headers={
        "Authorization": "Bearer " + os.environ["API_KEY"],
    },
    json={
        "model": "gpt-3.5-turbo-instruct",
        "prompt": "Write a short greeting:",
        "max_tokens": 64,
        "temperature": 0.7,
        "stream": False,
    },
)
print(response)

var client = java.net.http.HttpClient.newHttpClient();
var body = "{\"model\":\"gpt-3.5-turbo-instruct\",\"prompt\":\"Write a short greeting:\",\"max_tokens\":64,\"temperature\":0.7,\"stream\":false}";
var request = java.net.http.HttpRequest.newBuilder()
    .uri(java.net.URI.create("https://10000router.com/v1/completions"))
    .header("Authorization", "Bearer " + System.getenv("API_KEY"))
    .POST(java.net.http.HttpRequest.BodyPublishers.ofString(body))
    .build();
var response = client.send(request, java.net.http.HttpResponse.BodyHandlers.ofString());
System.out.println(response.body());

using System.Net.Http.Json;

using var client = new HttpClient();
client.DefaultRequestHeaders.Authorization = new("Bearer", Environment.GetEnvironmentVariable("API_KEY"));
var response = await client.PostAsJsonAsync(
    "https://10000router.com/v1/completions",
    new {
        model = "gpt-3.5-turbo-instruct",
        prompt = "Write a short greeting:",
        max_tokens = 64,
        temperature = 0.7,
        stream = false
    });
Console.WriteLine(await response.Content.ReadAsStringAsync());

返回响应

响应示例

{
  "id": "cmpl-abc123",
  "object": "text_completion",
  "created": 1710000000,
  "model": "gpt-3.5-turbo-instruct",
  "choices": [
    {
      "text": "Hello! How can I help you today?",
      "index": 0,
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 6,
    "completion_tokens": 8,
    "total_tokens": 14
  }
}

{
  "error": {
    "message": "Missing required parameter: prompt",
    "type": "invalid_request_error",
    "param": "prompt",
    "code": null
  }
}

{
  "error": {
    "message": "Rate limit reached for completions",
    "type": "rate_limit_exceeded",
    "param": null,
    "code": null
  }
}

{
  "error": {
    "message": "Invalid authentication credentials",
    "type": "invalid_request_error",
    "param": null,
    "code": null
  }
}

返回字段参数

响应字段按对象层级拆分为可折叠区块，默认展开；可收起暂时不关注的对象，减少长响应的视觉干扰。

流式响应

当请求设置 stream=true 时，响应的 Content-Type 为 text/event-stream，每个事件包含一个增量补全对象。客户端应按顺序拼接 choices[].text，并在收到 data: [DONE] 后结束读取：

data: {"id":"cmpl-abc123","object":"text_completion","created":1710000000,"model":"gpt-3.5-turbo-instruct","choices":[{"text":"Hello","index":0,"finish_reason":null}]}

data: {"id":"cmpl-abc123","object":"text_completion","created":1710000000,"model":"gpt-3.5-turbo-instruct","choices":[{"text":"! How can I help?","index":0,"finish_reason":"stop"}]}

data: [DONE]

流式响应中的每个 choices 元素只包含当前增量文本；usage 通常在非流式响应中返回。客户端应处理网络中断并在完成读取后释放连接。