Create response (OpenAI Responses API)

OpenAI Responses API, used to create model responses. Supports multi-turn conversations, tool calling, reasoning, and other features.

Add asynchronous image generation examples for OpenAI Responses API. background=true is used to submit background tasks; the create API returns status=queued and the full id immediately, then use that id to call GET /v1/responses/{response_id} to query the result.

Asynchronous image generation flow

1. Call POST /v1/responses, and pass tools=[{ "type": "image_generation" }] and background=true in the request body.
2. Save the full id returned by the create API, for example resp_xxx. Do not truncate, rewrite, or save only metadata.task_id.
3. Call GET /v1/responses/{response_id} to query until status=completed or status=failed.
4. After success, read output[].url to get the image URL.

Example of creating an asynchronous image task

POST /v1/responses
Content-Type: application/json

{
 "model": "gpt-image-2-2k",
 "input": "生成一张 2K 横屏产品海报，白色运动鞋，干净棚拍光线，高级商业摄影风格。",
 "stream": false,
 "tools": [
 {
 "type": "image_generation"
 }
 ],
 "background": true
}

Create API response example:

{
 "id": "resp_xxx",
 "created_at": 1780649146,
 "error": null,
 "incomplete_details": null,
 "instructions": null,
 "metadata": {
 "task_id": "010c31178d26436ca6194cde07931b33",
 "credits": "23"
 },
 "model": "gpt-image-2-2k",
 "object": "response",
 "output": [],
 "parallel_tool_calls": true,
 "temperature": null,
 "tool_choice": null,
 "tools": null,
 "top_p": null,
 "max_output_tokens": null,
 "previous_response_id": null,
 "reasoning": null,
 "status": "queued",
 "text": null,
 "truncation": null,
 "usage": null,
 "user": null,
 "store": null
}

Supported models: gpt-image-2-2k, gpt-image-2-4k, and nano-banana-pro.