OpenAI enables GIF creation in ChatGPT with Images 2.6 update

ChatGPT can now make GIFs. You can start from a written prompt or a photo you upload, and the chatbot returns a looping animation instead of a single still image.

The feature is rolling out alongside OpenAI’s new Images 2.6 model, according to The Verge. The internet’s favorite reaction format finally has a native home inside the world’s most famous chatbot.

What the GIF tool actually does

The workflow is simple on purpose. You describe what you want, or hand ChatGPT a picture, and it animates the result.

Once the GIF exists, you get some control over it. Users can change the animation speed, edit individual frames, and share the finished file with other people.

OpenAI’s own demo is about as on-brand for the internet as it gets. The prompt was “can you make a gif of a cat dancing?” ChatGPT produced the animation and offered options to adjust playback speed and step through it frame by frame.

Frame-level editing matters more than it sounds. It means you are not stuck rerolling the whole animation because one frame came out wrong. You can fix the weird paw and keep the rest of the dance.

Before this update, ChatGPT had no way to produce a GIF natively in one step. Users who wanted an animation had to generate a series of separate images, then stitch them together in an outside tool.

Under the hood: frames, flags, and a quiet code change

The groundwork showed up in OpenAI’s code before the consumer launch. On November 3, 2026, OpenAI merged a pull request in its Codex repository that enabled GIF output through the Images API.

That change supports animations with between 2 and 16 frames. The default setting is 9 frames, with each frame displayed for 200 milliseconds.

For non-engineers: a default GIF is a short loop of nine pictures, each flashing on screen for a fifth of a second. That is enough for a quick gesture or a simple motion, not a short film.

The code also supports spritesheet-based edits and keeps a still preview visible while the animation generates. A spritesheet is a single image that holds every frame laid out in a grid, a technique long used in video game animation.

On the developer side, the capability sits behind a feature flag that is currently disabled. It also has not yet been publicly documented in the API.

How we got here

This release builds on OpenAI’s previous image update. ChatGPT Images 2.5 launched on September 8, 2026, and improved image generation speed by up to 50%.

That version also improved editing fidelity and added features such as direct sketching and more intuitive editing options.

The broader GPT Image series succeeded DALL·E as OpenAI’s image engine. It now powers the native image tools in both ChatGPT and the API.

The series includes variants with different priorities. Flare is tuned for speed, while Sunburst is built for detail.

The GPT Image 2.6 model is also expected to bring more than animation. Anticipated additions include support for 4K resolution and markup-based edits, where users annotate an image to indicate changes.

What this means

The most immediate effect is on everyday creators. Social media managers, marketers, and meme makers live on short, punchy visuals, and GIFs remain a staple of that diet.

Developers are the group to watch next. The Images API code already supports GIF output, but the feature flag and missing documentation mean third-party apps cannot rely on it yet.

The technical limits also deserve a reality check. A ceiling of 16 frames keeps these animations short and simple. This is a tool for reaction loops and quick motion, not a replacement for video generation.