Image Generation

OpenAI image generation API reference: create images from prompts with dall-e-3 or gpt-image models, covering size, background, quality and response format.

Create images from a prompt using the native OpenAI format.

📌 Basics

Endpoint: POST /v1/images/generations/
Request domain: https://anyrouter.win
Authentication: Bearer Token
Content type: application/json

📥 Request Parameters (Request Body)

Parameter	Type	Required	Default	Description
model	`string`	No	`dall-e-2`	Model to use: `dall-e-2`, `dall-e-3`, or `gpt-image-1`, `gpt-image-1.5`, `gpt-image-2`.
prompt	`string`	Yes	-	Text description of the image. Up to 32k characters for GPT-image-1.
n	`integer`	No	`1`	Number of images (1-10). `dall-e-3` only supports 1.
size	`string`	No	`1024x1024`	Size. `gpt-image-1` supports `1024x1024`, `1536x1024`, `1024x1536`, or `auto`.
background	`string`	No	`auto`	Background transparency (`gpt-image-1` only): `transparent`, `opaque`, `auto`.
moderation	`string`	No	`auto`	Content moderation level: `low` (less restrictive) or `auto`.
quality	`string`	No	-	Quality of the generated image (HD/Standard).
style	`string`	No	-	Image style.

📤 Response

Status code: 200 OK

Field	Type	Description
created	`integer`	Unix timestamp when the request was created.
data	`array`	List of image objects.
└─ url	`string`	Hosted URL of the image.
└─ b64_json	`string`	Base64-encoded image data.
usage	`object`	Usage statistics.
└─ total_tokens	`integer`	Total tokens consumed by this request.

📝 Request Example

{
  "model": "gpt-image-1",
  "prompt": "A cute baby sea otter",
  "n": 1,
  "size": "1024x1024",
  "background": "transparent"
}

📝 Response Example

{
  "created": 1713833628,
  "data": [
    {
      "url": "https://example.com/generated_image.png",
      "b64_json": ""
    }
  ],
  "usage": {
    "total_tokens": 100,
    "input_tokens": 50,
    "output_tokens": 50,
    "input_tokens_details": {
      "text_tokens": 10,
      "image_tokens": 40
    }
  }
}

💡 Notes

Transparency support: When background is set to transparent, make sure the backend returns a format that supports an alpha channel, such as png or webp.
Model differences: Different models (DALL-E vs GPT-image) have different limits on size and n; choose a model based on your actual business needs.