Create an image

POST/v1/images/generations

Generate images from a text prompt.

Authorization

bearerAuth

AuthorizationBearer <token>

Use a platform API key in the format Authorization: Bearer your-api-key.

In: header

Request Body

application/json

TypeScript Definitions

Use the request body type in TypeScript.

model*string

Platform model ID.

prompt*string

Image generation or editing prompt, counted in Unicode characters excluding surrounding whitespace; cannot be blank.

Length1 <= length <= 32000

n?integer

Number of images to generate.

Range1 <= value <= 10

Default1

size?string

auto or WIDTHxHEIGHT, default auto. Custom area: 655,360–20,000,000 pixels; aspect ratio: 1:3–3:1. Omit aspect_ratio and area when specifying dimensions.

Default"auto"

aspect_ratio?string

Default auto, determined by the enhanced prompt and, for editing, reference images. Alternatively, use X:Y with integers X and Y from 1–300 and a ratio of 1:3–3:1. Only used when size is omitted or auto.

area?string

Area preset, default 4K. Only used when size is omitted or auto; actual dimensions are in data[].size.

Value in

"1K"
"2K"
"4K"

output_format?string

Image format. jpg and jpeg are equivalent; responses use jpeg.

Default"jpeg"

Value in

"jpeg"
"jpg"
"png"
"webp"

response_format?string

url links expire after 24 hours; b64_json returns Base64 without streaming. Total Base64 output area is limited to 20,000,000 pixels for PNG or 80,000,000 pixels for JPEG/WebP, summed across images.

Default"url"

Value in

"url"
"b64_json"

output_compression?integer

JPEG/WebP compression quality; higher values produce larger files. Ignored for PNG.

Range0 <= value <= 100

Default95

stream?boolean

Return each image over SSE; only supports response_format set to url.

Defaultfalse

aigc_watermark?boolean

Add a visible AIGC watermark; does not affect the implicit identifier in file metadata.

Defaultfalse

disable_prompt_enhanced?boolean

Disable prompt enhancement. When disabled, only approximately the first 4,096 tokens apply; specify a concrete size or fixed aspect_ratio.

Defaultfalse

Response Body

`application/json`

type*string

created?integer

Event Unix timestamp.

model?string

Actual model ID used.

image_index*integer

Image index, starting at 0.

Range0 <= value

url*string

Temporary link to the generated image.

size?string

Actual image WIDTHxHEIGHT.

output_format?string

Actual output format.

Value in

"jpeg"
"png"
"webp"

revised_prompt?string

Prompt used for this image.

curl -X POST "https://example.com/v1/images/generations" \
  -H "Authorization: Bearer your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "MODEL_ID",
    "prompt": "Generate a photo of a green desk lamp on a wooden table, with a light-colored wall and soft natural lighting.",
    "size": "2048x2048",
    "n": 1,
    "response_format": "url"
  }'

{
  "created": 1794355200,
  "data": [
    {
      "url": "https://example.com/lamp-result.jpg",
      "size": "2048x2048",
      "revised_prompt": "A green desk lamp sits on a wooden table in front of a light-colored wall, with soft natural light illuminating the shade and tabletop."
    }
  ],
  "output_format": "jpeg",
  "request_id": "REQUEST_ID"
}

Set the client timeout to at least 600 seconds. Generation may continue after a timeout or disconnection; sending again creates a new request.

MiniMax image generation APISee request parameters and response structure.

Authorization

Request Body

Response Body

200application/json

400application/json

401application/json

402application/json

422application/json

429application/json

Streaming response events

`application/json`

`application/json`

`application/json`

`application/json`

`application/json`