perplexity-sonar-reasoning-pro-ssvip 多模态推理 API 使用文档
基于 Perplexity Sonar Reasoning Pro 模型（perplexity-sonar-reasoning-pro-ssvip）的 Responses 端点接口，在视觉理解与联网搜索的基础上融入链式推理能力，答题前会先输出完整的思考过程（reasoning）再给出结论。相比 Sonar Pro，该接口更适合需要逐步分析、结合搜索结果推理的复杂视觉问答场景，返回结果同时包含推理过程（reasoning/reasoning_details）与 URL 引用注释（annotations）。
🌐 请求地址
https://www.dmxapi.cn/v1/responsesWARNING
请妥善保管您的 API Key！严禁将密钥泄露给他人、硬编码到代码中或提交到公开的代码仓库。如果怀疑密钥已泄露，请立即前往 DMXAPI 官网重新生成。
🤖 模型名称
perplexity-sonar-reasoning-pro-ssvip
📋 多模态推理 示例代码
python
import requests
import json
# 步骤1: 配置 API 连接信息
# DMXAPI 服务端点地址
url = "https://www.dmxapi.cn/v1/responses"
# DMXAPI 密钥 (请替换为您自己的密钥)
api_key = "sk-******************************************"
# 步骤2: 配置请求头
headers = {
# 指定请求体为 JSON 格式
"Content-Type": "application/json",
# token 认证方式，直接传入 API Key（无需加 Bearer 前缀）
"Authorization": f"{api_key}",
}
# 步骤3: 配置请求参数
payload = {
# 【model】(string, 必填) 指定调用的模型名称
# 使用 perplexity-sonar-reasoning-pro-ssvip 启用视觉理解、联网搜索与链式推理能力
"model": "perplexity-sonar-reasoning-pro-ssvip",
# 【input】(array, 必填) 对话输入内容，支持多模态混合数组
# 每个元素为一条消息，包含 role 和 content 字段
"input": [
{
# 【role】(string, 必填) 消息角色
# 可选值: "user"(用户消息) / "assistant"(模型回复) / "system"(系统提示)
"role": "user",
# 【content】(array, 必填) 消息内容，支持文本与图像混合输入
"content": [
{
# 【type】(string, 必填) 内容类型为纯文本
# "input_text" 表示文本输入，对应用户的提问文字
"type": "input_text",
# 【text】(string, 必填) 针对图像提出的问题或指令
"text": "What is in this image?"
},
{
# 【type】(string, 必填) 内容类型为图像 URL
# "image_url" 表示通过公网 URL 传入图像，模型将对其进行视觉理解
"type": "image_url",
# 【image_url】(object, 必填) 图像地址对象
"image_url": {
# 【url】(string, 必填) 图像的公网可访问地址
# 支持常见图片格式（JPEG、PNG、WebP、GIF 等）
"url": "https://live.staticflickr.com/3851/14825276609_098cac593d_b.jpg"
}
}
]
}
]
}
# 步骤4: 发送请求并输出结果
# 发送 POST 请求到 API 服务器
response = requests.post(url, headers=headers, json=payload)
# 格式化输出 JSON 响应
# indent=2: 缩进 2 空格，便于阅读
# ensure_ascii=False: 正确显示中文字符
print(json.dumps(response.json(), indent=2, ensure_ascii=False))📄 返回示例
json
{
"id": "gen-1775222825-xcRIplilrfVZtivtsuHv",
"object": "chat.completion",
"created": 1775222825,
"model": "perplexity-sonar-reasoning-pro-ssvip",
"choices": [
{
"index": 0,
"logprobs": null,
"finish_reason": "stop",
"native_finish_reason": "stop",
"message": {
"role": "assistant",
"content": "The image shows **bottlenose dolphins** swimming in ocean water, captured in what appears to be a close-up or near-surface view.[1] Multiple dolphins are visible moving through the blue-green water, with their distinctive curved dorsal fins and sleek bodies characteristic of the species.[4][5]",
"refusal": null,
"reasoning": "The user is asking me to describe what is in the image they've provided. Looking at the image, I can see dolphins in ocean water. The image shows what appears to be bottlenose dolphins swimming/moving through blue-green ocean water with whitecaps/waves visible.\n\nI should describe what's visible in the image: dolphins (appearing to be bottlenose dolphins) in ocean water with visible waves.",
"reasoning_details": [
{
"type": "reasoning.text",
"text": "The user is asking me to describe what is in the image they've provided. Looking at the image, I can see dolphins in ocean water...",
"format": "unknown",
"index": 0
}
],
"annotations": [
{
"type": "url_citation",
"url_citation": {
"url": "https://www.livenowfox.com/news/watch-spectacular-video-shows-dolphins-leaping-out-water-off-san-diego-coast",
"title": "Watch: 'Spectacular' video shows dolphins leaping out of water off ...",
"start_index": 0,
"end_index": 0
}
},
{
"type": "url_citation",
"url_citation": {
"url": "https://www.istockphoto.com/photos/dolphin-jumping",
"title": "12592 Dolphin Jumping Images and Stock Photos - iStock",
"start_index": 0,
"end_index": 0
}
},
{
"type": "url_citation",
"url_citation": {
"url": "https://www.gettyimages.com/photos/dolphin-jumping",
"title": "2658 Dolphin Jumping Stock Photos, High-Res Pictures, and Images",
"start_index": 0,
"end_index": 0
}
}
]
}
}
],
"usage": {
"total_tokens": 906,
"input_tokens": 0,
"input_tokens_details": {
"cached_tokens": 0
},
"output_tokens": 906,
"output_tokens_details": {
"reasoning_tokens": 0
}
}
}返回字段说明
| 字段 | 说明 |
|---|---|
id | 本次请求的唯一 ID |
object | 固定为 chat.completion |
model | 实际调用的模型名称 |
choices[].message.content | 模型对图像内容的最终分析结论，引用来源以 [n] 标注 |
choices[].message.reasoning | 模型作答前的完整推理过程文本，包含搜索结果的整合分析 |
choices[].message.reasoning_details | 推理过程的结构化详情数组，type 固定为 reasoning.text |
choices[].message.annotations | URL 引用注释列表，对应正文中 [n] 标注的来源链接 |
choices[].finish_reason | 停止原因，stop 表示正常结束 |
usage.total_tokens | 本次请求消耗的总 token 数 |
© 2026 DMXAPI perplexity-sonar-reasoning-pro-ssvip 多模态推理