ComfyUI Node
GPT Image 2.6 Sunburst (BudgetPixel)
OpenAI GPT Image 2.6 Sunburst — the detailed GPT Image model, for dense text, complex layouts and precise edits that keep everything you did not ask to change. Text-to-image, and editing or combining up to 16 reference images; transparent backgrounds; output up to 4K. New in 2.6: complexity sets how much the model plans before it renders. Billing: per output image by quality, the same at every size and with any number of references — low 70, medium 120, high 175, xhigh 340, max 675 credits. Default quality medium (120 credits). Pricing: 120 credits per generation.
GPT Image 2.6 Sunburst (BudgetPixel)
- reference_images
- images
◄prompt►
◄aspect_ratio1:1►
◄backgroundauto►
◄complexityhigh►
◄num_images1►
◄output_formatpng►
◄qualitymedium►
◄resolution1K►
◄force_rerunfalse►
CategoryBudgetPixel/Image
Inputs (10)
| Name | Type | Default | Description |
|---|---|---|---|
| prompt | STRING | Text description of the image to generate. Write `@handle` to cast a character: a BudgetPixel AI influencer or one of your own ([`GET /v1/characters`](/api-reference/characters/list-characters-you-can-cast)). | |
| aspect_ratioopt | COMBO | 1:1 | Output aspect ratio. |
| backgroundopt | COMBO | auto | Output background. transparent returns a cut-out with a transparent background, from text or an edit (as PNG, or WebP if asked: JPEG has no transparency). auto lets the model decide. |
| complexityopt | COMBO | high | How much the model plans before it renders. high suits dense text, complex layouts and scenes where objects relate to each other; medium and low plan less, are faster, and suit focused edits such as changing a background or a sign. Does not change the price. |
| num_imagesopt | INT | 11–4 | Number of images to generate. |
| output_formatopt | COMBO | png | Output image format. |
| qualityopt | COMBO | medium | Rendering detail. Per image, at any size: low 70, medium 120, high 175, xhigh 340, max 675 credits. |
| reference_imagesopt | IMAGE | Optional reference images (up to 16) to edit or combine into one image. Each item is a public image URL, a data URI, raw base64, or an uploaded-file URL from POST /v1/uploads. Omit for text-to-image. References do not change the price. | |
| resolutionopt | COMBO | 1K | Output size tier; the price is the same at every size. 1:1 renders 1024x1024 at 1K and 2048x2048 at 2K or 4K. 16:9 and 9:16 render 2048x1152 / 1152x2048 at 1K or 2K, and 3840x2160 / 2160x3840 at 4K. 3:2, 2:3, 4:3 and 3:4 have one size each (1536x1024, 1024x1536, 1536x1152, 1152x1536). With aspect_ratio auto the model picks the size, and an edit keeps its input's shape. |
| force_rerunopt | BOOLEAN | false | — |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| images | IMAGE | — |