Choose an iteration pace
Flare is designed for quick iteration; Sunburst emphasizes detail and consistency. Compare actual latency and output on your own prompts.
ChatGPT Images 2.6
An OpenAI image model guide
Explore Flare and Sunburst, GIF generation, markup-guided edits, SVG output, and complexity control. Compare supplied examples before choosing a workflow.
Key features
Flare and Sunburst
The API offers gpt-image-2.6-flare and gpt-image-2.6-sunburst. Choose the model that fits your speed and visual-quality priorities.
Animated GIF generation
Create a looping GIF with the API generation endpoint. Frame count and frame duration control the animation; GIF output is generation-only.
Markup-guided edits
A transparent PNG overlay can identify the region to change. Match the overlay dimensions to the first input image and describe the intended edit.
SVG output
Request output_format=svg for scalable artwork such as logos and marks. SVG output is supported on the generation endpoint.
Complexity control
Set complexity to low, medium, or high to control planning effort. It is separate from quality, and the documented default is high.
Reference-led workflows
Use reference images to communicate a character, object, composition, or visual direction. Check the limits and controls of the product you use.
Layout for professional visuals
Explore posters, product imagery, interface concepts, and other composed visuals. Check text, alignment, and detail before publishing.
Review before publishing
Review generated text, identities, edits, and image metadata. Consult the current model documentation for supported formats and limitations.
Technical specifications
API models
Two variants: gpt-image-2.6-flare and gpt-image-2.6-sunburst.
Speed / detail
Compare Flare for iteration and Sunburst for detail on your workload.
GIF + SVG
Generation-only output formats; GIF frame count ranges from 2 to 64.
complexity
low · medium · high; default high, separate from quality.
markup
Transparent PNG overlay matching the first input image dimensions.
Flare / Sunburst
Use the exact model identifier supported by your chosen API endpoint.
$5 / $8 / $30
Per 1M API tokens: text input / image input / image output.
Check availability
Open Studio shows the current SugarCube model catalog.
Use cases
Candid photography
A supplied portrait example with natural-looking texture and lighting. Inspect skin, clothing, and fine details at full size.
Campaign ads
A supplied streetwear campaign example for the brand “Thread,” illustrating the placement of a short tagline.
Comic reels
A supplied vertical four-panel narrative example for exploring visual pacing and storyboards.
Interface previews
A supplied mobile app concept showing hierarchy, spacing, and an interface composition.
Process infographics
A supplied coffee-machine diagram with labeled stages and a bean-to-brew reading order.
Product cutouts
A supplied product-cutout example for inspecting edges, isolation, and background transparency.
Deck slides & charts
A supplied pitch-style slide with market-size diagrams and a clear information hierarchy.
Fast photoreal drafts
A supplied photorealistic draft for comparing composition, texture, and detail with the other portrait example.
Room & object edits
A supplied kitchen scene illustrating a reference-led room and object-editing workflow.
Markup edit · vase recolor
Supplied kitchen photo, transparent circle markup, and two recolored-vase examples. Use the sequence to inspect whether stems, shape, and surrounding details remain consistent.
Input · kitchen photo
Markup preview
Flare · cobalt vase
Sunburst · cobalt vase
Complexity · paper-cut championship
Three supplied paper-cut examples labeled low, medium, and high complexity. Compare trophy detail, lettering, and composition.
complexity=low
complexity=medium
complexity=high · default
SVG · Field & Flour bakery logo
Supplied Field & Flour vector examples, labeled Flare and Sunburst, shown on a light board. SVG output is available for API generations.
Prompt examples
Looping GIF · monkey lantern
A little monkey in an olive-green hooded coat stands in a snowy pine forest at dusk…
Show full prompt
A little monkey in an olive-green hooded coat stands in a snowy pine forest at dusk. A wooden cabin glows in the background. The monkey holds a glowing amber paper lantern at chest height, slowly lifts it, looks up and blinks, then lowers it back to its starting pose. Keep the lantern in its hands and the camera still. Make the motion loop smoothly.
SVG / logo · Field & Flour
Create an original, non-infringing logo for a company called Field & Flour, a local bakery…
Show full prompt
Create an original, non-infringing logo for a company called Field & Flour, a local bakery. The logo should feel warm, simple, and timeless. Use clean, vector-like shapes, a strong silhouette, and balanced negative space. Favor simplicity over detail so it reads clearly at small and large sizes. Flat design, minimal strokes, no gradients unless essential. Fully transparent background. Deliver a single centered logo with generous padding, clean alpha edges, and no solid backdrop, scenery, checkerboard, or watermark.
Complexity compare · paper-cut
A premium handcrafted layered paper-cut illustration of a football team celebrating a championship…
Show full prompt
A premium handcrafted layered paper-cut illustration of a football team celebrating a championship victory inside a packed stadium.
Markup edit · vase recolor
Recolor the striped vase inside the red circle to solid cobalt blue ceramic…
Show full prompt
Recolor the striped vase inside the red circle to solid cobalt blue ceramic. Keep its shape, the dried stems, and the rest of the kitchen unchanged. Remove the red circle from the finished photo.
Exact text · campaign ad
Streetwear campaign still with brand name and a single exact tagline for layout control.
Show note
Create a streetwear campaign still for a fictional brand called Thread. Render the tagline “Yours to Create.” exactly once. Keep the copy legible and the composition uncluttered.
Infographic · coffee machine
Process diagram with labeled stages — check the labels, not just the look.
Show note
Use clear stage labels and a single reading order. Inspect small type, arrows, and connector lines at the final display size.
What is ChatGPT Images 2.6?
ChatGPT Images 2.6 is the subject of this editorial guide to OpenAI image generation and editing. The API documentation describes two model variants, GIF generation, SVG output, markup-guided edits, and a separate complexity control. Capabilities can differ between ChatGPT and API endpoints.
Developers can choose gpt-image-2.6-flare or gpt-image-2.6-sunburst. Flare is oriented toward faster iteration; Sunburst emphasizes visual quality and consistency across edits. Test the same prompt, references, size, and quality to compare them fairly.
Use the examples on this page to examine the kinds of work you need: a portrait, campaign still, comic, interface concept, infographic, product cutout, or presentation slide. The supplied artwork illustrates these workflows; it is not a record of paid tasks run through SugarCube.
For motion, the API generation endpoint supports GIF output with frame controls. For targeted changes, supply a transparent PNG markup overlay alongside the first input image. SVG generation produces vector artwork. The complexity setting controls planning effort separately from the quality setting.
- Inspect text, faces, brand details, composition, and unintended changes after each generation or edit. If a region must remain pixel-identical, use compositing and verify the final file.
- Check current OpenAI documentation for model availability, output support, and limitations. SugarCube Studio lists only the models currently integrated in this product.
For animation, describe one small action, a fixed camera, and a clear return to the starting pose.
Prompting note · looping motion
For local edits, state both what should change and what should stay the same, then inspect the surrounding region.
Prompting note · controlled edits
Flare vs Sunburst
Choose a starting point, then compare the same prompts, references, size, and quality. Real response time and output quality depend on the request.
GPT-Image-2.6 Flare Fast
An option to test when fast iteration is the main priority.
- Compare response time on your own workload
- Use low quality when a draft is sufficient
- Inspect poster layouts and product details
- GIF and SVG output use the generation endpoint
- Keep the prompt and settings fixed for a fair comparison
GPT-Image-2.6 Sunburst Quality
An option to test when visual detail and edit consistency matter most.
- Compare detail at the size you plan to publish
- Inspect consistency over successive edits
- Review small text, diagrams, and exact brand details
- Use supported markup and complexity controls
- Compare Flare once the output meets your quality needs
What to compare with Images 2.5
Speed
Compare actual response time with identical inputs and output settings.
GIF
Create animated GIFs on the generation endpoint with output_format=gif and frame controls.
Markup
Supply a transparent PNG markup overlay for a targeted image edit.
SVG
Request vector output on the generation endpoint with output_format=svg.
Complexity
Compare low, medium, and high planning effort separately from quality.
References
Keep the same reference images when comparing generations or edits.
Layout
Inspect typography, spacing, product geometry, and visual hierarchy.
Models
The model IDs are gpt-image-2.6-flare and gpt-image-2.6-sunburst.
Pricing & availability
The table shows documented OpenAI API token rates for Flare and Sunburst. These are API prices, separate from SugarCube credits and subscription plans; check the current provider pricing before use.
| Token type | Per 1M tokens | Cached |
|---|---|---|
| Text input | $5 | $1.25 |
| Image input | $8 | $2 |
| Image output | $30 | — |
API rates, separate from SugarCube credits
The listed rates apply to text input, image input, and image output for the two API variants. SugarCube Studio shows its own supported models and applicable credit cost before a task starts.
OpenAI API
Flare
Sunburst
Check current rates
Studio catalog varies
FAQs
What does this ChatGPT Images 2.6 page cover?
This is an editorial guide to image generation and editing workflows, supplied visual examples, and documented API controls for gpt-image-2.6-flare and gpt-image-2.6-sunburst. ChatGPT and API capabilities can differ.
Where can I check current availability?
Check the current OpenAI documentation for ChatGPT and API availability. Open SugarCube Studio to see the model catalog currently integrated here; Images 2.6 is not currently listed.
Should I choose Flare or Sunburst?
Compare Flare when iteration speed matters and Sunburst when detail or edit consistency matters most. Use identical prompts, references, size, and quality, then inspect the result for your task.
How do GIF parameters work in the API?
GIF output is supported on the generation endpoint. Set output_format=gif. frame_count ranges from 2–64, with a default of 9; frame_duration_ms ranges from 20–2000, with a default of 200. GIF output is not supported on the edit endpoint.
What does the complexity parameter control?
complexity accepts low, medium, or high, with high as the default. It controls planning effort before rendering and is separate from quality. Compare it while keeping your other settings fixed.
Can the API generate SVG output?
The generation endpoint supports output_format=svg, including transparent backgrounds. SVG output is not supported on the edit endpoint. Inspect the resulting paths and appearance before production use.
How does markup-guided editing work?
Provide a transparent PNG overlay that matches the first input image dimensions, identify the region to change, and describe the edit in the prompt. Inspect both the changed area and the surrounding image.
What are the documented API token rates?
For either API variant, per 1M tokens: text input $5, cached text input $1.25, image input $8, cached image input $2, and image output $30. These provider rates are separate from SugarCube credits; confirm the current pricing before use.
What should I inspect in generated images?
Check spelling, labels, faces, character or product identity, layout, and unintended edit changes. Use compositing when a region must remain pixel-identical. The examples on this page are supplied illustrations, not SugarCube performance benchmarks.
Can I use Images 2.6 in SugarCube Studio today?
Images 2.6 is not currently listed in SugarCube Studio. This page is an editorial model briefing. The Open Studio link opens the existing workspace so you can explore the models that are available now.
Explore your next image workflow
Take a prompt idea into SugarCube Studio and choose from the models available there. Images 2.6 is covered here as an editorial guide and is not currently listed in Studio.