ChatGPT will now make you animated GIFs

OpenAI’s latest rollout, Images 2.6, finally lets ChatGPT spit out animated GIFs without the hacky frame‑by‑frame juggling users have been piecing together for months. A single prompt – or an uploaded photo – now returns a looping animation that you can slow down, speed up, or edit frame by frame right inside the chat window. The change arrives alongside a claimed jump to over 4 billion images generated each week across ChatGPT and the API, a 33 percent rise from the 3 billion reported two months earlier.

The technical leap rests on the new model’s ability to keep visual consistency across a short sequence. Earlier versions, such as Images 2.5, could produce a series of stills, but the character or product often drifted in style between frames, making a smooth loop impossible. Images 2.6, however, preserves key attributes – color palette, shape, and lighting – long enough for a handful of frames to stitch together, as OpenAI demonstrated with a “cat dancing” GIF that stays recognizably the same cat from start to finish. Users can then slide a speed control or tap individual frames to swap a detail, turning a rough sketch into a quick‑look animation for social media, a demo, or an internal presentation.

For developers, the update is more than a UI tweak. OpenAI is exposing two API variants: GPT‑Image‑2.6 Flare, tuned for raw speed (as fast as six seconds on low‑quality settings), and GPT‑Image‑2.6 Sunburst, which favors higher‑fidelity output and tighter cross‑frame consistency. This mirrors the split introduced with Images 2.5, where Adobe’s Firefly adopted a similar “speed vs. quality” dichotomy. The documentation still lists the 2.5 model IDs, a lag that explainx.ai flagged as a reminder not to hard‑code the 2.6 identifiers until they appear in the official model list.

The headline figure – more than 4 billion images per week – is eye‑catching, but the breakdown is vague. OpenAI’s press release bundles static pictures, GIFs, and API calls together, offering no ratio for how many of those images are animated loops. Moreover, the September count of 3 billion used a similar “all‑image” definition, making a tidy one‑third growth claim plausible but not rigorously comparable. Analysts therefore treat the number as a gauge of overall platform activity rather than a precise metric of GIF adoption. What is clearer is the strategic timing: the GIF feature lands months after OpenAI shuttered its dedicated video tool, Sora, in March, signalling a pivot toward short‑form motion that can be delivered at the scale of a text‑based chat.

For marketers, designers, and support teams the practical payoff is immediate. A product manager can upload a static mock‑up of a new device, ask ChatGPT to animate a button press, adjust the playback speed, and embed the resulting GIF directly into a product spec without ever leaving the chat. A social media manager can generate a meme‑style reaction GIF on the fly, tweak a single frame for brand color, and share it across platforms. The integration also lowers the barrier for non‑technical creators who previously needed separate stitching tools or third‑party services.