Gemini 3 Pro Image 图像生成

OmniRouters Google Gemini 图片协议参考 for the gemini-3-pro-image model.

接口

POST /v1beta/models/gemini-3-pro-image:generateContent

最小请求示例

curl https://omnirouters.com/v1beta/models/gemini-3-pro-image:generateContent \
  -H "x-goog-api-key: $OMNIROUTERS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{\n  "contents": [\n    {\n      "role": "user",\n      "parts": [\n        { "text": "Create a cinematic image of a flower in a glasshouse." }\n      ]\n    }\n  ],\n  "generationConfig": {\n    "responseModalities": ["TEXT", "IMAGE"]\n  }\n}'

此接口根据文本提示词生成图片。

请求体

Field	Type	必填	Description
contents	array	必填	Gemini content 数组，包含文本或图片 Part。
generationConfig	object	否	生成配置；图片输出通常设置 responseModalities。

Gemini image responses are returned in candidates[].content.parts[].inlineData.

返回参数

{
  "candidates": [
    {
      "content": {
        "role": "model",
        "parts": [
          { "inlineData": { "data": "...", "mimeType": "image/png" } }
        ]
      },
      "finishReason": "STOP"
    }
  ]
}

说明

Model-specific validation and output availability depend on the selected model and current OmniRouters routing configuration. For shared routing guidance, see Omni-Image API.