> ## Documentation Index
> Fetch the complete documentation index at: https://phaseo.app/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# 按价格、延迟或吞吐量排序提供商

> 使用请求级提供商排序，为某个模型优先选择最便宜、最快或吞吐量最高的路由。

当模型保持不变，但你希望网关为某个请求以不同方式排列提供商报价时，可以使用此方案。

## 目标

* 为某个请求优先选择最便宜的提供商。
* 为交互式流程优先选择延迟最低的提供商。
* 批量生成时优先选择吞吐量最高的提供商。

## 1. 在请求中添加 `provider.sort`

对于文本请求，请直接在 `provider` 对象中设置排序方式。

<CodeGroup>
  ```bash cURL theme={null}
  curl https://api.phaseo.app/v1/responses \
    -H "Authorization: Bearer YOUR_API_KEY" \
    -H "Content-Type: application/json" \
    -d '{
      "model": "google/gemini-3.1-flash-lite",
      "input": "Give me one release-note bullet for the last deploy.",
      "provider": {
        "sort": "price"
      }
    }'
  ```

  ```typescript TypeScript SDK theme={null}
  import Phaseo from "@phaseo/sdk";

  const client = new Phaseo({ apiKey: process.env.PHASEO_API_KEY! });

  const response = await client.generateResponse({
    model: "google/gemini-3.1-flash-lite",
    input: "Give me one release-note bullet for the last deploy.",
    provider: {
      sort: "price",
    },
  });

  console.log(response.output_text);
  ```

  ```python Python SDK theme={null}
  from phaseo import Phaseo

  client = Phaseo(api_key="YOUR_API_KEY")

  response = client.generate_response(
      {
          "model": "google/gemini-3.1-flash-lite",
          "input": "Give me one release-note bullet for the last deploy.",
          "provider": {
              "sort": "price",
          },
      }
  )

  print(response.get("output_text"))
  ```

  ```go Go SDK theme={null}
  package main

  import (
    "context"
    "fmt"

    phaseo "github.com/phaseoteam/Phaseo/packages/sdk/sdk-go/v3"
  )

  func main() {
    client := phaseo.New("YOUR_API_KEY", "https://api.phaseo.app/v1")

    response, err := client.GenerateResponse(context.Background(), phaseo.ResponsesRequest{
      Model: "google/gemini-3.1-flash-lite",
      Input: "Give me one release-note bullet for the last deploy.",
      Provider: map[string]interface{}{
        "sort": "price",
      },
    })
    if err != nil {
      panic(err)
    }

    fmt.Println(response)
  }
  ```

  ```csharp C# SDK theme={null}
  using PhaseoSdk;
  using System.Collections.Generic;

  var client = new Phaseo("YOUR_API_KEY");

  var response = await client.GenerateResponse(new Dictionary<string, object>
  {
      ["model"] = "google/gemini-3.1-flash-lite",
      ["input"] = "Give me one release-note bullet for the last deploy.",
      ["provider"] = new Dictionary<string, object>
      {
          ["sort"] = "price"
      }
  });

  Console.WriteLine(response);
  ```

  ```php PHP SDK theme={null}
  <?php
  require 'vendor/autoload.php';

  use Phaseo\Sdk\Phaseo;

  $client = new Phaseo(getenv('PHASEO_API_KEY') ?: 'YOUR_API_KEY');

  $response = $client->generateResponse([
      'model' => 'google/gemini-3.1-flash-lite',
      'input' => 'Give me one release-note bullet for the last deploy.',
      'provider' => [
          'sort' => 'price',
      ],
  ]);

  print_r($response);
  ```

  ```ruby Ruby SDK theme={null}
  require 'phaseo_sdk'

  client = PhaseoSdk::Phaseo.new(api_key: ENV.fetch('PHASEO_API_KEY', 'YOUR_API_KEY'))

  response = client.generate_response(
    model: 'google/gemini-3.1-flash-lite',
    input: 'Give me one release-note bullet for the last deploy.',
    provider: {
      sort: 'price'
    }
  )

  puts response
  ```
</CodeGroup>

此流程支持以下路由排序方式：

* `price`
* `latency`
* `throughput`

## 2. 选择适合工作负载的排序方式

用法：

* 当成本比尾延迟更重要时使用 `price`
* 聊天、辅助工具和人工参与的工具使用 `latency`
* 高并发生成或回填任务使用 `throughput`

## 3. 了解网关的比较依据

当你明确指定请求级排序时，网关会以确定性方式排列候选项，而不是使用常规的平衡加权随机排序。

对于文本模型：

* `price` 会在符合条件的提供商之间比较统一的价格基准
* 如果存在通用文本计量项，网关会优先选择 `input_text_tokens` 和 `output_text_tokens` 匹配的提供商
* `latency` 使用最新的提供商延迟数据
* `throughput` 使用最新的吞吐量测量值

## 4. 保持合理的的提供商范围

必要时先缩小候选范围，再进行排序，效果会更好。

例如，只在一组已批准的提供商中进行排序：

<CodeGroup>
  ```bash cURL theme={null}
  curl https://api.phaseo.app/v1/responses \
    -H "Authorization: Bearer YOUR_API_KEY" \
    -H "Content-Type: application/json" \
    -d '{
      "model": "google/gemini-3.1-flash-lite",
      "input": "Summarize the incident in one sentence.",
      "provider": {
        "only": ["google-vertex", "google-vertex-eu"],
        "sort": "latency"
      }
    }'
  ```

  ```typescript TypeScript SDK theme={null}
  import Phaseo from "@phaseo/sdk";

  const client = new Phaseo({ apiKey: process.env.PHASEO_API_KEY! });

  const response = await client.generateResponse({
    model: "google/gemini-3.1-flash-lite",
    input: "Summarize the incident in one sentence.",
    provider: {
      only: ["google-vertex", "google-vertex-eu"],
      sort: "latency",
    },
  });

  console.log(response.output_text);
  ```

  ```python Python SDK theme={null}
  from phaseo import Phaseo

  client = Phaseo(api_key="YOUR_API_KEY")

  response = client.generate_response(
      {
          "model": "google/gemini-3.1-flash-lite",
          "input": "Summarize the incident in one sentence.",
          "provider": {
              "only": ["google-vertex", "google-vertex-eu"],
              "sort": "latency",
          },
      }
  )

  print(response.get("output_text"))
  ```

  ```go Go SDK theme={null}
  package main

  import (
    "context"
    "fmt"

    phaseo "github.com/phaseoteam/Phaseo/packages/sdk/sdk-go/v3"
  )

  func main() {
    client := phaseo.New("YOUR_API_KEY", "https://api.phaseo.app/v1")

    response, err := client.GenerateResponse(context.Background(), phaseo.ResponsesRequest{
      Model: "google/gemini-3.1-flash-lite",
      Input: "Summarize the incident in one sentence.",
      Provider: map[string]interface{}{
        "only": []string{"google-vertex", "google-vertex-eu"},
        "sort": "latency",
      },
    })
    if err != nil {
      panic(err)
    }

    fmt.Println(response)
  }
  ```

  ```csharp C# SDK theme={null}
  using PhaseoSdk;
  using System.Collections.Generic;

  var client = new Phaseo("YOUR_API_KEY");

  var response = await client.GenerateResponse(new Dictionary<string, object>
  {
      ["model"] = "google/gemini-3.1-flash-lite",
      ["input"] = "Summarize the incident in one sentence.",
      ["provider"] = new Dictionary<string, object>
      {
          ["only"] = new[] { "google-vertex", "google-vertex-eu" },
          ["sort"] = "latency"
      }
  });

  Console.WriteLine(response);
  ```

  ```php PHP SDK theme={null}
  <?php
  require 'vendor/autoload.php';

  use Phaseo\Sdk\Phaseo;

  $client = new Phaseo(getenv('PHASEO_API_KEY') ?: 'YOUR_API_KEY');

  $response = $client->generateResponse([
      'model' => 'google/gemini-3.1-flash-lite',
      'input' => 'Summarize the incident in one sentence.',
      'provider' => [
          'only' => ['google-vertex', 'google-vertex-eu'],
          'sort' => 'latency',
      ],
  ]);

  print_r($response);
  ```

  ```ruby Ruby SDK theme={null}
  require 'phaseo_sdk'

  client = PhaseoSdk::Phaseo.new(api_key: ENV.fetch('PHASEO_API_KEY', 'YOUR_API_KEY'))

  response = client.generate_response(
    model: 'google/gemini-3.1-flash-lite',
    input: 'Summarize the incident in one sentence.',
    provider: {
      only: ['google-vertex', 'google-vertex-eu'],
      sort: 'latency'
    }
  )

  puts response
  ```
</CodeGroup>

## 5. 验证排序结果

调试时，请在 **Gateway -> 用量** 中检查请求并查找：

* 已考虑的提供商
* 已排序的提供商
* 价格、延迟或吞吐量的路由评分因素

这样可以轻松确认网关是否按预期排序。

## 相关指南

* [路由和回退](../guides/routing-and-fallbacks.mdx)
* [推出预设并调试路由](./preset-rollout-and-routing-debug.mdx)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.