ChatGPT’s GIF Upgrade Turns AI Into a Meme Factory—Minus the Crypto Chop 🐱

OpenAI is rolling out animated GIF generation in ChatGPT, powered by a new image-generation model called Images 2.6 that the company has not formally introduced through its usual product announcements. The feature reportedly lets users create animations by describing them in plain English or uploading an existing photo and asking ChatGPT to bring it to life.

The generated GIF interface includes controls for changing playback speed, editing the animation and inspecting individual frames. OpenAI demonstrated the capability with the request: “can you make a gif of a cat dancing?” The Verge reported that the feature was unavailable on some accounts during testing, while a separate interaction produced a series of frames without assembling them into an animation.

The rollout appears gradual and is expected to reach ChatGPT and Codex users across web, desktop and mobile platforms. Developers are also receiving two new API models: GPT-Image-2.6 Flare, designed for faster generation, and GPT-Image-2.6 Sunburst, built for higher-quality results. OpenAI has not published a standalone announcement about the update, and its performance could not be independently tested.

The release follows the September 8 launch of ChatGPT Images 2.5, which introduced sharper details, richer textures, more natural lighting and more precise editing. OpenAI says image-generation latency dropped by as much as 50% compared with Images 2.0. GPT-Image-2.5 Flare and GPT-Image-2.5 Sunburst are also available through the API, with Flare serving as the fast default and Sunburst focused on premium editing precision.

In previous testing, Images 2.5 competed with Google’s Nano Banana 2 across six categories, with each model winning three. OpenAI’s model delivered strong illustration quality and reduced oversharpening seen in earlier versions, although it continued to make mistakes when rendering text. The GIF feature also arrives months after OpenAI shut down Sora, its standalone AI video-generation app, in March.