Sample Request

curl --location 'https://cloud.leonardo.ai/api/rest/v2/generations' \
     --header 'authorization: Bearer {api-key}' \
     --header 'Content-Type: application/json' \
     --data '{
       "public": false,
       "model": "openai/gpt-image-2.6-flare",
       "parameters": {
           "quality": "HIGH",
           "complexity": "HIGH",
           "prompt": "Koala with purple hat",
           "quantity": 2,
           "width": 1376,
           "height": 768,
           "prompt_enhance": "OFF",
           "style_ids": [
               "111dc692-d470-4eec-b791-3475abac4c46"
           ],
           "guidances": {
               "image_reference": [
                   {
                       "image": {
                         "id": "00000000-0000-0000-0000-000000000001",
                         "type": "UPLOADED"
                       }
                   }
               ]
           }
        }
     }'

Recipe

API Request Endpoint, Headers, Parameters

Endpoint

https://cloud.leonardo.ai/api/rest/v2/generations

Headers

--header "accept: application/json" \
--header "authorization: Bearer <YOUR_API_KEY>" \
--header "content-type: application/json"

Body Parameters

Parameter	Type	Definition
annotation	`boolean`	Optional. Set to `true` when an image reference has marks drawn on it that show what to change.
complexity	`string`	Optional. Controls how much the model plans the image before it renders, separate from `quality`. Higher values help with dense text and complex layouts, and can take longer. Accepts `LOW`, `MEDIUM`, or `HIGH`. Defaults to `HIGH`.
guidances	`object`	Optional. Object defining image guidance inputs that influence the generated result.
guidances.image_reference	`array`	Optional. Array of up to 16 reference images. Each reference includes image.id, image.type (`UPLOADED`, `GENERATED`, `VARIATION`, `URL`, or `BASE64`), and an optional order. Use image.url when type is `URL`, and image.data (base64-encoded image bytes) when type is `BASE64`. Note: GPT Image 2.6 Flare does not use `strength` for image references.
height	`integer`	Optional. Height input resolution. Defaults to `1024`. Must be a multiple of `16` between `16` and `3840`. The longer side can be at most three times the shorter side. Refer to the section below for common values.
model	`string`	Required. Model identifier. Set to `openai/gpt-image-2.6-flare`.
prompt	`string`	Required. Text prompt describing what image you want the model to generate. Maximum 9999 characters.
prompt_enhance	`string`	Optional. Controls prompt enhancement. Accepts `AUTO`, `ON`, or `OFF`. Defaults to `AUTO`.
public	`boolean`	Optional. Boolean flag that determines whether the generation is public (`true`) or private (`false`).
quality	`string`	Optional. Accepts `LOW`, `MEDIUM`, `HIGH`, `XHIGH`, or `MAX`. Defaults to `HIGH`.
quantity	`integer`	Optional. Number of images to generate in a single request (1-8). Defaults to `4`.
style_ids	`array`	Optional. Array of style UUIDs used to apply preset artistic styles to the output (limited to 1 style). Defaults to `111dc692-d470-4eec-b791-3475abac4c46` (Dynamic). Refer to the table below.
transparency	`boolean`	Optional. Set to `true` to allow the image to be generated with a transparent background.
width	`integer`	Optional. Width input resolution. Defaults to `1024`. Must be a multiple of `16` between `16` and `3840`. The longer side can be at most three times the shorter side. Refer to the section below for common values.

Aspect Ratio Settings

Aspect Ratio	Width	Height
1:1	1024	1024
2:3	848	1264
3:2	1264	848
16:9	1376	768
9:16	768	1376

Resolution Constraints

Multiples of 16 only
width and height must each be a multiple of 16, from 16 up to 3840.
Maximum 3:1 ratio
The longer side can be at most three times the shorter side.
No auto sizing
0 is not an accepted width or height value, and size: "auto" is not supported and must not be passed to the API.

List of Style IDs

Preset Style	UUID
3D Render	`debdf72a-91a4-467b-bf61-cc02bdeb69c6`
Acrylic	`3cbb655a-7ca4-463f-b697-8a03ad67327c`
Cinematic	`a5632c7c-ddbb-4e2f-ba34-8456ab3ac436`
Creative	`6fedbf1f-4a17-45ec-84fb-92fe524a29ef`
Dynamic	`111dc692-d470-4eec-b791-3475abac4c46`
Fashion	`594c4a08-a522-4e0e-b7ff-e4dac4b6b622`
Game Concept	`09d2b5b5-d7c5-4c02-905d-9f84051640f4`
Graphic Design 2D	`703d6fe5-7f1c-4a9e-8da0-5331f214d5cf`
Graphic Design 3D	`7d7c2bc5-4b12-4ac3-81a9-630057e9e89f`
Illustration	`645e4195-f63d-4715-a3f2-3fb1e6eb8c70`
None	`556c1ee5-ec38-42e8-955a-1e82dad0ffa1`
Portrait	`ab5a4220-7c42-41e5-a578-eddb9fed3d75`
Portrait Cinematic	`4edb03c9-8a26-4041-9d01-f85b5d4abd71`
Portrait Fashion	`0d34f8e1-46d4-428f-8ddd-4b11811fa7c9`
Pro B&W photography	`22a9a7d2-2166-4d86-80ff-22e2643adbcf`
Pro Color Photography	`7c3f932b-a572-47cb-9b9b-f20211e63b5b`
Pro Film Photography	`581ba6d6-5aac-4492-bebe-54c424a0d46e`
Ray Traced	`b504f83c-3326-4947-82e1-7fe9e839ec0f`
Stock Photo	`5bdc3f2a-1be6-4d1c-8e77-992a30824a2c`
Vibrant	`dee282d3-891f-4f73-ba02-7f8131e5541b`
Watercolor	`1db308ce-c7ad-4d10-96fd-592fa6b75cc4`