Tasks and inputs
GPT-Image-2.6 Sunburst is listed for Image-text-to-image, Image-to-image, Text-to-image. It accepts text, image and produces image. Use the publisher’s input format and preprocessing for your chosen deployment.
SyncDev model directory · Profile updated Nov 17, 2026
GPT-Image-2.6 Sunburst is OpenAI's detail-first image generation and editing model, adding SVG, animated GIF output and markup-guided edits.
The specification score summarizes recorded features and limits. It is not a benchmark result.
Pricing conditions: Indicative token pricing per 1M: text input $5 (cached $1.25); image input $8 (cached $2); image output $30. No per-image list price. Half-price Batch image rates are published, but Batch is marked unsupported for this model; test before relying on it.
Recorded modality support shows what this model can accept and produce. An undocumented modality is shown as unknown rather than assumed to be unsupported.
Accepts
Produces
Most image models look fine on the first render. The trouble starts on the fourth edit, when a label has blurred, a hand has grown a finger and the logo has drifted two shades off brand. GPT-Image-2.6 Sunburst is OpenAI's answer for that kind of iterative, detail-heavy work: the variant of the 2.6 family tuned to keep compositions intact while you keep changing them.
Its sibling, GPT-Image-2.6 Flare, is tuned for speed instead. Both shipped on 4 November 2026 and cost exactly the same per token, so the choice between them is about fidelity versus turnaround, never price.
The release is less about raw quality than about new output types and controls:
SVG is the quiet headline. A designer who can open the result in a vector editor and nudge a path saves a round-trip through tracing tools.
| Field | Value |
|---|---|
| API ID | gpt-image-2.6-sunburst (single snapshot) |
| Released | 4 November 2026 |
| Input | Text, image |
| Output | Image (raster, SVG, animated GIF) |
| Endpoints | v1/images/generations, v1/images/edits only |
| Text input | $5 per 1M tokens ($1.25 cached) |
| Image input | $8 per 1M tokens ($2 cached) |
| Image output | $30 per 1M tokens |
It is not served on Responses or Chat Completions. If your application generates images as a tool call inside a chat flow, you will need a separate call to the Images endpoints.
One date conflict is worth flagging: some third-party listings show 28 October 2026, while OpenAI dates the release to 4 November. OpenAI's date is the one to trust for when the model became callable.
OpenAI publishes no per-image list price for this model. Cost depends on how many tokens the input images and the output consume, which in turn depends on size and the complexity setting. Run a representative batch of your real prompts and read the usage object on each response; that is the only reliable per-image figure.
Batch is murkier than it should be. OpenAI publishes half-price batch rates for image tokens, yet also marks the Batch endpoint as unsupported for this model. Before planning an overnight job around the discount, submit a small batch and see whether it is accepted. All of these figures are indicative; check your own account.
This is the model OpenAI points to when retiring older image models. gpt-image-1.5, gpt-image-1-mini and chatgpt-image-latest all shut down on 1 December 2026, with Sunburst or Flare as the named replacement. gpt-image-2 follows on 10 February 2027. If you still call GPT-Image-1.5 or GPT-Image-2, the deadline is close.
Coming from GPT Image 2.5 Sunburst, token prices are unchanged, so the decision rests on whether SVG, GIF and the complexity control earn the switch.
Take three real briefs: a product shot with text on the packaging, a diagram with labels, and a brand illustration. Run each through five rounds of edits on Sunburst and on Flare. Score what survives: legible text, consistent colour, unchanged elements you did not ask to change. If Flare holds up on your material, take the speed. If it drifts by round three, Sunburst is the one to pay for in time.
More image models sit in the text-to-image directory.
Compare the model on the work you need to complete.
GPT-Image-2.6 Sunburst is listed for Image-text-to-image, Image-to-image, Text-to-image. It accepts text, image and produces image. Use the publisher’s input format and preprocessing for your chosen deployment.
The listed route has starting rates of $5.00 input and $30.00 output per million tokens. Choose the output size and quality your design needs. Include drafts and revisions when budgeting an accepted image.
Use the model identifier and account requirements for the selected API route. Confirm its request limits and output format with one complete task before connecting a larger workflow.
These records share a directory category. Check the source card, output type and deployment requirements before treating a model as a replacement.
recraft
Recraft V4.1 Flash V2 is Recraft's fastest, cheapest image model: 1K raster output in about 1.1 s for $0.007, quality close to Recraft V4.1.
MiniMax
MiniMax H3 Image is an open-weight 33.1B image generation and editing model producing native ~16MP output in six steps, with up to 16 reference images.
MiniMax
MiniMax H3 Image Highspeed is the faster hosted tier of MiniMax's H3 Image generator and editor, at double the image-token price of the standard tier.
OpenAI
GPT-Image-2.6 Flare is OpenAI's fast image generation and editing model for interactive iteration, with SVG, animated GIF and markup-guided edits.
krea
Krea 3 is Krea's in-house image model for text-to-image and instruction-based edits at 1K or native 2K, priced from $0.03 per image.
Alibaba
Qwen-Image-2.1-Turbo creates and edits images, including reference-guided and transparency workflows; hosted billing is separate from local deployment.