OpenAI has released two new image models in its API, and ChatGPT is gaining animated GIFs. OpenAI's API changelog entry for Nov 4, 2026 says it "Released GPT-Image-2.6 Sunburst and GPT-Image-2.6 Flare for image generation and editing, with adjustable complexity, SVG generation, and animated GIF generation."

The ChatGPT side comes from The Verge, which reported on Nov 4, 2026: "ChatGPT can now create GIFs using a prompt or a photo uploaded to the AI chatbot." OpenAI has no announcement page for the ChatGPT rollout, and its ChatGPT release notes had no entry for it when we checked on Nov 5, 2026.

#What are Sunburst and Flare?

OpenAI splits the two models by use:

  • Sunburst. Its model page describes "Image generation and editing for detailed visual work" and says: "Use it for detailed compositions and image editing workflows."
  • Flare. Its model page describes "Image generation and editing for fast iteration" and says: "Use it for interactive image generation and editing workflows."

Both take text and image inputs and output images. Both work with the image generation endpoint v1/images/generations and the image edit endpoint v1/images/edits. The model pages list the Responses endpoint as not supported.

OpenAI's image generation guide says: "Both models support complexity control, markup-guided edits, SVG output, and animated GIF output through the Image API."

#What can the API models do?

  • Complexity. "Use the complexity request field to control how much the image model plans before rendering," OpenAI's guide says. The levels are low, medium and high, and both models default to high. OpenAI warns: "Higher complexity can increase latency and introduce unwanted changes, so compare results for your task." It adds: "Adjusting complexity has no separate fee."
  • Animated GIFs. Developers set output_format to gif on the generations endpoint. The guide allows 2 to 64 frames, with a default of 9, and 20 to 2,000 milliseconds per frame, rounded to 10 ms, with a default of 200. "GIF output is for image generation; the image edit endpoint does not support GIF output."
  • SVG. OpenAI says to "Use SVG for graphics that need editable paths, shapes, and colors, such as icons and illustrations." As with GIFs, "The image edit endpoint does not support SVG output, and an SVG is not a supported raster input for image edits."
  • Markup-guided edits. The changelog says OpenAI "Added markup-guided editing for supported GPT Image 2.5 and 2.6 models." The guide says: "Use an annotation layer to show where a change should happen and describe the change in prompt."

OpenAI lists limits for GIFs: "GIF output does not support streaming, partial images, custom output_compression, or prompt_rewriter."

#How much does it cost?

The two model pages list the same prices, as of Nov 5, 2026:

Type

Price per 1M tokens

Text input

$5

Cached text input

$1.25

Image input

$8

Cached image input

$2

Image output

$30

The guide says "Image output costs $30 per million billable tokens" and "Markup-guided edits use the same token rates as other edits, with no extra fee."

#What changes in ChatGPT?

All of the following comes from The Verge's report. The Verge reported that the GIF update "is rolling out alongside OpenAI’s new Images 2.6 model, allows you to adjust the GIF’s animation speed, edit individual frames, and share it with others."

The Verge wrote: "Before this change, ChatGPT couldn’t natively generate GIFs in a single step." It described an example OpenAI shared, in which ChatGPT answers the prompt "can you make a gif of a cat dancing?" with a GIF and options to change the playback speed and view it frame by frame.

On availability, The Verge reported: "Images 2.6 is rolling out to all ChatGPT and Codex users across web, desktop, and mobile apps." No source names plan-level limits for GIFs.

Two further claims are OpenAI's, according to The Verge:

  • Speed. The Images 2.6 update "also enables faster image generation in ChatGPT, OpenAI says."
  • Volume. "The company says users now generate over 4 billion images per week across ChatGPT and its API, a jump from the 3 billion per week it reported in September."

No source says whether ChatGPT's Images 2.6 runs on Sunburst or Flare.

For OpenAI's previous API model launch, see our story on GPT-6.1 Terra, a cheaper model for computer use. For other ChatGPT and Codex changes, see ChatGPT and Codex moving to a shared usage allowance.

Our analysis: The API release is well documented: two models with a clear split, specific GIF limits and published prices. The consumer side is not. Everything about GIFs in ChatGPT, including who gets them and the faster generation claim, rests on one outlet's report of what OpenAI showed it. For developers, the practical limits matter: GIFs and SVGs come only from the generations endpoint, and GIF output drops streaming and partial images. Higher complexity is free but can add latency, so the default of high is worth testing against lower settings.

#What is unclear?

  • ChatGPT model. Which API model, if either, powers Images 2.6 in ChatGPT.
  • Plan limits. Any GIF limits by ChatGPT plan.
  • Official notes. OpenAI's own release note for the ChatGPT rollout.
  • Rate limits. OpenAI says only that rate limits vary by usage tier and account.