ChatGPT Images 2.6 turns prompts and photos into animated GIFs

A new image model brings one-step GIFs with frame editing, faster pictures and two new API models.

OpenAI is rolling out Images 2.6, a new image model that lets ChatGPT create animated GIFs from a written prompt or an uploaded photo. The update was reported on November 4, 2026. Users can change the playback speed and edit single frames, and OpenAI says image generation also gets faster.

Key facts

OpenAI is rolling out animated GIF creation in ChatGPT together with a new image model, Images 2.6, first reported by The Verge on November 4, 2026.
A GIF can start from a text prompt or from an uploaded photo. After it is made, users can change its speed, view and edit single frames, and share it.
The rollout covers all ChatGPT and Codex users on the web, desktop and mobile apps.
Developers get two API models: GPT-Image-2.6 Flare for speed and GPT-Image-2.6 Sunburst for higher quality.
OpenAI says people now make more than 4 billion images a week across ChatGPT and its API, up from 3 billion in September.

What happened

Until this update, ChatGPT could not produce a GIF in one step. People who wanted a moving image had to generate frames one by one and stitch them together somewhere else. With Images 2.6, a single request is enough.

In its demo, OpenAI used the prompt “Can you make a gif of a cat dancing?”. The model produces a series of slightly different versions of the same character and strings them together like a flipbook. The Decoder notes that this only works because the model keeps the character consistent from one frame to the next, something earlier image models struggled with.

Once the animation is ready, ChatGPT offers controls to speed it up or slow it down and to step through the frames, so a single frame can be fixed without starting over. OpenAI also says Images 2.6 makes ordinary image generation in ChatGPT faster.

For developers, the company is releasing GPT-Image-2.6 Flare, aimed at efficient generation, and GPT-Image-2.6 Sunburst, aimed at quality. The Next Web points out that this two model split was introduced with Images 2.5 in September. It also recalls that OpenAI shut down Sora, its AI video app, in March, which makes GIFs the main way to get moving pictures inside ChatGPT for now.

Why it matters

GIFs are one of the most common formats on the internet: reactions in group chats, small animations on product pages, short loops in newsletters and presentations. Making them used to mean a separate app, a stack of frames and some patience. Putting the whole process inside ChatGPT turns it into a quick request that anyone can try.

Frame editing is the practical detail. With image generators, the usual problem is that one bad result forces a full restart. Being able to correct a single frame keeps the parts that already work.

The usage numbers show why OpenAI keeps updating its image tools. More than 4 billion images a week, a billion more than in September, means pictures are now one of the main things people do with ChatGPT. These figures come from OpenAI and cover all image use, not only GIFs. The pace is also fast: Images 2.0, the first OpenAI image model that reasons before drawing, arrived in April, and 2.5 followed in September.

How to try it

The new feature works well with photos you already have, and the clearer the starting image, the better the animation. A good first step is a clean portrait of yourself. Our free Professional profile photo prompt turns a photo of you into a neat headshot while keeping your real face, and that picture is a solid base for a short animated loop.

Open the prompt and fill in the fields: what your photo shows, where you will use it, the outfit and the background. For example: photo “Me from the chest up, taken near a window, neutral expression”, use “LinkedIn profile”, outfit “Dark blazer over a light shirt”, background “Plain light grey”. Paste the result into ChatGPT together with your own photo.

When you have the headshot, you can ask in the same chat: “Make a short GIF of this photo where I smile and wave, keep my face exactly the same”. Then use the speed control and look at each frame. Check that your face, hair and clothes stay the same in every frame, and fix any frame that drifts. Use only photos of yourself or of people who agreed. The feature is still rolling out, so it may not appear in every account yet.

Try it

Professional profile photo from your own photo

Frequently asked questions

Can ChatGPT make GIFs now?

Yes. With the Images 2.6 update, ChatGPT can create an animated GIF from a text prompt or an uploaded photo in one step. After it is made you can change the speed, edit single frames and share it.

Who can use GIF creation in ChatGPT?

OpenAI says the feature is rolling out to all ChatGPT and Codex users on the web, desktop and mobile apps. Because it is a gradual rollout, some accounts may see it later than others.

What are GPT-Image-2.6 Flare and Sunburst?

They are the two new API models for developers. Flare is built for speed and efficiency, Sunburst for higher quality images. OpenAI first split its image models this way with Images 2.5 in September.

Sources