ChatGPT Images 2.6: animated GIFs straight from the chat
ChatGPT Images 2.6 makes animated GIFs from a prompt or a photo, remembers characters and exports SVG, and the API gets Flare and Sunburst. What it can do.
This article in other languages:العربيةDeutschEspañolFrançaisItaliano日本語Polski简体中文
In this article
On November 4, OpenAI released ChatGPT Images 2.6, a new version of its image generation model. The headline feature is animated GIFs made in ChatGPT in a single step, from a text description or an uploaded photo. The model is coming to ChatGPT and Codex on the web, desktop and mobile apps, and the API gets two new models: GPT-Image-2.6 Flare and GPT-Image-2.6 Sunburst.
What’s new in ChatGPT Images 2.6?
Until now, ChatGPT couldn’t make a GIF in one step. Now all it takes is a prompt like “can you make a gif of a cat dancing?” (the example OpenAI used in its demo, as Decrypt reports) or a photo you want to bring to life.
According to launch coverage, Images 2.6 brings, among other things:
- animated GIFs from text or a photo, with adjustable playback speed, a preview, editing of individual frames and sharing of the finished file,
- saved reference images, a library of characters, pets, mascots and people you can reuse in later images,
- marking an area of the image that should change,
- SVG export, so you get vector graphics instead of just pixels,
- faster image generation in ChatGPT (according to OpenAI).
Interestingly, GIFs arrive a few months after OpenAI shut down Sora, its AI video app, in March. An animated GIF is of course a much simpler format than video, but motion is back in ChatGPT.
Flare and Sunburst: Images 2.6 in the API
For developers, OpenAI repeats the split from September’s Images 2.5: one model for speed, another for quality.
- GPT-Image-2.6 Flare is built for generation speed: interactive interfaces, high request volumes, apps where response time matters.
- GPT-Image-2.6 Sunburst focuses on quality and consistency across successive edits: product photos, characters, images that get revised many times.
The announcement includes no new pricing (at least according to AlphaSignal). We also don’t know the limits on GIF length, frame count or resolution.
Why is everyone talking about Images 2.6?
Because it’s a feature anyone can show their friends right away, and the scale is huge. According to OpenAI, ChatGPT and the API already generate more than 4 billion images a week, up from 3 billion in September (that figure covers all images, not just GIFs).
The pace is impressive too: Images 2.0 in April was OpenAI’s first image model that “thinks” before drawing, Images 2.5 came in September, and now, less than two months later, we have 2.6 with animation.
How to try ChatGPT Images 2.6
Just ask ChatGPT (or Codex) for a GIF, e.g. “make a GIF of a dancing cat”, or upload a photo and describe how it should move. GIFs don’t cost extra, but your account’s usual limits apply, and ChatGPT and Codex draw from a shared usage limit.
What to watch out for
First, the rollout pace. The rollout is gradual and not everyone has GIFs yet. Decrypt didn’t get access on its accounts, and in its test the model produced a series of frames instead of an animation (if you don’t see GIFs yet, try again in a few days).
Second, reference images. A library of saved characters and people is convenient, but OpenAI hasn’t yet explained in detail how long it keeps them or how to delete them. Think twice before uploading photos of other people.
Third, the lack of independent tests. So far the promise of faster generation comes only from OpenAI, so let’s take it with a grain of salt until the first comparisons appear.
FAQ
What is ChatGPT Images 2.6?
It’s a new version of OpenAI’s image generation model in ChatGPT and Codex, released on November 4, 2026. Its main new feature is animated GIFs from text or a photo, plus reference images, marking areas to change and SVG export.
Are GIFs in ChatGPT free?
According to launch coverage, GIFs don’t cost extra, and Images 2.6 is rolling out to ChatGPT and Codex users, with the usual account limits. The rollout is gradual, so the feature may not have reached you yet.
What’s the difference between GPT-Image-2.6 Flare and Sunburst?
Both are Images 2.6 models available in the API. Flare is optimized for generation speed, and Sunburst for image quality and consistency across successive edits.