ChatGPT can now produce animated GIFs. Type a prompt or upload a photo, and the chatbot returns a looping animation that you can slow down, speed up or tweak frame by frame before sharing it. OpenAI is rolling the feature out with Images 2.6, its latest image model.
Until now, ChatGPT could not make a GIF in one step. OpenAI’s example is the prompt “can you make a gif of a cat dancing?”, which gives you the animation along with playback controls and a frame-by-frame view.
The Decoder explains what changed: Images 2.6 draws the same character consistently several times, and those near-identical pictures are played back like the pages of a flipbook. With older models a character’s look would change too much between frames, which is why GIFs were out of reach.
The update reaches all ChatGPT and Codex users on the web, desktop and mobile apps, according to OpenAI. The company also says Images 2.6 generates ordinary pictures faster than the previous model, so the speed gain is not limited to GIFs.
Developers get two API models alongside it, GPT-Image-2.6 Flare, the efficient option, and Sunburst, the higher-quality option. Both names were first used with the previous Images 2.5 release.
Sharing is built into the tool, so a finished animation can be passed on from the app itself, and OpenAI puts the rollout across web, desktop and phone apps.
The company also gave a usage figure: people now make more than four billion images a week through ChatGPT and the API, up from three billion in September.
