Image Generation (Sync)

In addition to the async task API from the previous page, GPT Image also supports OpenAI-compatible synchronous generation: the request waits for the image to finish generating and returns the result directly in the response, so no task polling is required.

Pick either of the two sync styles:

• Chat Completions style: POST /v1/chat/completions
• Images API style: POST /v1/images/generations

Style 1: Chat Completions (POST /v1/chat/completions)

Call it like a Chat Completions request with model set to gpt-image; authentication works the same as for regular chat models. Clients already wired to OpenAI Chat Completions can reuse the integration as-is.

This style also supports image-to-image: in the example panel below, switch between the Text-to-image / Image-to-image request bodies. For image-to-image, set the content of the user message to an array and pass the reference image (URL or Base64 data URI) in an image_url part — matching OpenAI multimodal input.

curl -X POST "https://www.omwai.xyz/v1/chat/completions" \  -H "Content-Type: application/json" \  -d '{    "model": "gpt-image",    "stream": false,    "messages": [      {        "role": "user",        "content": "一只可爱的猫，赛博朋克风格，霓虹灯背景"      }    ],    "size": "1024x1024",    "quality": "high"  }'

{
  "id": "chatcmpl-01J8ABCDEFGHIJKLMNOPQRST",
  "object": "chat.completion",
  "created": 1788256630,
  "model": "gpt-image",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "string"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 0,
    "completion_tokens": 0,
    "total_tokens": 0
  }
}

{
  "code": "string",
  "message": "string"
}

Style 2: Images API (POST /v1/images/generations)

Call it like the OpenAI Images API with model set to gpt-image. Supported parameters match the async endpoints (prompt, n, size, quality, background, output_format, moderation, etc.). When generation completes, the data array directly contains the final images (url or b64_json fields).

curl -X POST "https://www.omwai.xyz/v1/images/generations" \  -H "Content-Type: application/json" \  -d '{    "model": "gpt-image",    "prompt": "a cute cat, cyberpunk style, neon background",    "n": 1,    "size": "1024x1024",    "quality": "high"  }'

{
  "created": 1788256630,
  "data": [
    {
      "url": "string",
      "b64_json": "string",
      "revised_prompt": "string"
    }
  ],
  "usage": {
    "total_tokens": 0,
    "input_tokens": 0,
    "output_tokens": 0,
    "input_tokens_details": {
      "text_tokens": 0,
      "image_tokens": 0
    }
  }
}

{
  "code": "string",
  "message": "string"
}