POST /api/v1/images/generations with model: gpt-image-2-6-sunburst and a prompt to create one PNG. Add up to 16 reference images to edit, select quality and resolution, and retrieve the result using img_uuid.GPT Image 2.6 Sunburst API for Image Generation and Editing
Use
Get API Key
Generate, Edit and Control Rendering
Choose output quality
Start with low quality at 1K, then select medium, high, xhigh or max. The 1K low tier with up to 1,024 UTF-8 prompt bytes and no references costs 1.92 tokens ($0.0384).
Create transparent PNGs
Set
background: transparent to retain an alpha channel for cut-outs and layered layouts. Transparency does not add a charge.
Edit reference images
Pass up to 16 direct image URLs as
image_list: [{"url":"https://…"}]. Describe what to change and preserve. Reference dimensions are measured for input pricing.
model and prompt are required; quality and image_size together set the price.
GPT Image 2.6 Sunburst Request Fields and Limits
model and prompt
Use
gpt-image-2-6-sunburst and a non-empty prompt of up to 32,000 characters.Output controls
image_size: 1k, 2k or 4k. quality: low, medium, high, xhigh or max. Defaults are 1k and low. One PNG is returned.Ratio and background
Choose one of 15 documented ratios or auto. Auto uses 1:1 for text generation and follows the first reference for editing. Background can be opaque (default), transparent or auto.
Reference images
Up to 16 direct public HTTP(S) image URLs, at most 50MB each, returning HTTP 200 or 206 without redirects. Images must have readable dimensions.
Planning intensity
complexity: low, medium or high (default). Independent of rendering quality, with no extra charge.Unsupported options
Auto quality, custom pixel sizes, multiple outputs, GIF, SVG, masks and markup are not supported by this endpoint.
Fixed prepaid output and input tariffs. The tiers below include up to 1,024 UTF-8 prompt bytes, without references.
GPT Image 2.6 Sunburst API Pricing
low
1k: 1.92 tokens ($0.0384), 2k: 3.54 tokens ($0.0708), 4k: 5.65 tokens ($0.1130).
medium
1k: 3.12 tokens ($0.0624), 2k: 5.99 tokens ($0.1198), 4k: 9.73 tokens ($0.1946).
high
1k: 3.71 tokens ($0.0742), 2k: 7.19 tokens ($0.1438), 4k: 11.72 tokens ($0.2344).
xhigh
1k: 6.33 tokens ($0.1266), 2k: 12.51 tokens ($0.2502), 4k: 20.58 tokens ($0.4116).
max
1k: 13.82 tokens ($0.2764), 2k: 27.74 tokens ($0.5548), 4k: 45.88 tokens ($0.9176).
Additional prompt input
Each additional started 1,024 UTF-8 bytes of trimmed prompt text adds 0.33 tokens ($0.0066).
Reference input
Each reference image adds 2.1 tokens ($0.042) per started 1,048,576 pixels. Round up each image separately and sum.
Quote and refund
Call
POST /api/v1/pricing/quote with the same model and payload, including prompt and image_list. Input dimensions are measured before charging. Failed jobs refund their charge.Integrate the GPT Image 2.6 Sunburst API
2
2. Quote your request
POST to
/api/v1/pricing/quote with model and payload containing prompt, image_list and output settings. 3
3. Submit one generation
POST a flat request body with
model: gpt-image-2-6-sunburst to /api/v1/images/generations. Store the returned img_uuid. 4
4. Retrieve the PNG
Poll
GET /api/v1/jobs/detail using img_uuid or supply custom_callback_url. Download and store the image after success.Features where an edit should change what was asked and little else.
What Developers Build with GPT Image 2.6 Sunburst
Targeted photo edits
Apply instructions like recoloring an object or replacing a label to user photos where edit fidelity matters most.
Product cut-outs
Produce transparent PNG product shots for catalogs, marketplaces and compositing.
Multi-image compositions
Combine up to 16 references, such as products, logos supplied by the user and backgrounds, into one image.
Cost-tiered pipelines
Draft at
low for 0.32 tokens ($0.0064) at 1k, then re-render approved images at high or above.Editing, quality tiers and pricing.
GPT Image 2.6 Sunburst API FAQ
What is the endpoint?
POST https://api.goenhance.ai/api/v1/images/generations with model: gpt-image-2-6-sunburst.
What is the starting price?
Low / 1K with up to 1,024 UTF-8 prompt bytes and no references costs 1.92 tokens ($0.0384). Longer prompts and reference images add input charges.
What does an editing request cost?
Low / 1K, a short prompt and one 1024×1024 reference costs 4.02 tokens ($0.0804), before any account discount.
What are the defaults?
Low quality, 1K, opaque background and high planning complexity. Text generation with auto ratio uses 1:1.
Can I generate GIF or SVG files?
This endpoint currently returns one PNG per request. GIF, SVG, masks, markup and auto quality are not supported.
How many reference images can I use?
Up to 16 images. Their dimensions are measured independently for billing; transparency and planning intensity add no charge.