OpenAI Images 2.6: GIFs in ChatGPT, Two New API Image Models

OpenAI Images 2.6 adds GIF creation and saved reference images in ChatGPT, plus Sunburst and Flare API models with SVG, GIF and markup edits. What changed.

By PromptWises Editorial Team Published Updated 3 min read
Watch the 1-minute video summary on the PromptWises YouTube channel

PromptWises is reader-supported. Some links may earn us a commission at no extra cost to you. This never affects our ratings or recommendations.

OpenAI released Images 2.6 on November 4, 2026. In ChatGPT it means faster image generation, better consistency across edits, GIF creation and saved reference images you can reuse with @name. In the API it brings two new models, gpt-image-2.6-sunburst and gpt-image-2.6-flare, with a new complexity setting, SVG and animated GIF output, and markup-guided editing. OpenAI has not published prices for the new models or said which ChatGPT plans get the new features.

OpenAI’s own summary, from its news feed, is that Images 2.6 “brings faster image generation, GIF creation, more precise editing, and new API controls across ChatGPT, Codex, and the API.” The full announcement page returned an error to our fetcher, so this post relies on the API changelog, the image generation guide and the ChatGPT release notes.

What changes in ChatGPT

The ChatGPT release notes for November 4 list three changes:

  • Faster, more consistent images. Generation is quicker, and the model is better at keeping details the same across edits. OpenAI’s example is a character whose outfit stays the same when you change the background.
  • GIF creation (rolling out). ChatGPT can make looping animations, including from an uploaded sketch, with adjustable playback speed, and you can share the result.
  • Saved reference images. You can save a photo under a name and pull it into a later chat with @name instead of uploading it again. That is useful for a product shot, a logo or a recurring character.

The notes do not say which plans get GIF creation or saved references, or how many you can make. Image tools are already uneven across plans: OpenAI’s help centre says the Free no-ads option comes without image tools, and image generation with Thinking starts on Plus. Our ChatGPT review covers what each plan includes.

What changes for developers

The API changelog entry adds both models to v1/images/generations and v1/images/edits. The guide gives the details:

GPT Image 2.6 SunburstGPT Image 2.6 Flare
Model IDgpt-image-2.6-sunburstgpt-image-2.6-flare
Positioned forDetailed visual workFast iteration
complexitylow, medium, high (default high)low, medium, high (default high)
SVG and animated GIF outputYesYes
Markup-guided editsYesYes

Complexity controls how much the model plans before rendering. It has no separate fee, but OpenAI warns that higher complexity can increase latency and introduce unwanted changes, so it is worth testing medium against the default.

Markup-guided editing lets you draw on a transparent PNG layer the same size as the image to show what should change. It works on supported GPT Image 2.5 and 2.6 models through the edits endpoint. The markup applies to the first input image, and any other images act as references. The image and annotation layer can each be up to 16,777,216 pixels, the annotation file up to 50 MB, and markup cannot be combined with mask.

The guide also notes that you may need to complete API Organization Verification before using GPT Image models.

Update (11 November 2026): on November 10 OpenAI deprecated the older gpt-image-2 model and its gpt-image-2-2026-04-21 snapshot. Both are removed from the API on February 10, 2027, and OpenAI names gpt-image-2.6-sunburst or gpt-image-2.6-flare as replacements (OpenAI deprecations page).

What we do not know yet

  • Prices. Neither the changelog nor the guide lists per-image prices for the 2.6 models at the time of writing.
  • Plans. No plan list or limits for GIF creation and saved references in ChatGPT.
  • Codex. The summary mentions Codex, but the pages we could read give no Codex-specific details.
  • GIF limits. The guide confirms GIF and SVG output but gives no frame, length or size limits.

Should you care?

If you already make images in ChatGPT, saved reference images are the practical win: consistent product shots or characters without re-uploading every time. GIFs are handy for social posts and slides, but check the result frame by frame before you publish.

If you build on the API, try Flare for drafts and Sunburst for finals, and test the complexity setting rather than leaving it on high. Wait for published prices before you move production traffic. For how ChatGPT’s image tools compare with rivals, see ChatGPT vs Claude vs Gemini and our Gemini review. More launches are on our news page.

Frequently asked questions

What is OpenAI Images 2.6?

Images 2.6 is OpenAI's image generation update released on November 4, 2026. OpenAI describes it as faster image generation, GIF creation, more precise editing and new API controls across ChatGPT, Codex and the API. Developers get two new models, gpt-image-2.6-sunburst and gpt-image-2.6-flare.

Can ChatGPT make GIFs now?

Yes, GIF creation is rolling out in ChatGPT with Images 2.6, according to OpenAI's ChatGPT release notes. It makes looping animations, including from an uploaded sketch, with adjustable playback speed. OpenAI has not said which plans get it or whether there are limits.

What is the difference between GPT Image 2.6 Sunburst and Flare?

OpenAI's image generation guide says Sunburst is for detailed visual work and Flare is for fast iteration. Both support the complexity setting, markup-guided edits, SVG output and animated GIF output through the Image API.

How much does GPT Image 2.6 cost?

OpenAI's API changelog and image generation guide did not list per-image prices for the 2.6 models when we checked on November 9, 2026. The guide only says that adjusting the complexity setting has no separate fee. Check OpenAI's pricing page before you budget.

What are saved reference images in ChatGPT?

Saved reference images let you save a photo under a name and call it in a later chat with @name, so you do not have to upload it again. OpenAI lists the feature alongside Images 2.6 in its November 4 release notes, without plan details.

Keep reading