ChatGPT gains editable animated GIFs
OpenAI adds animated GIF generation in ChatGPT with Images 2.6. Users can start from a prompt or photo, then adjust speed and edit frames. Deployment also covers Codex and API.
ChatGPT can now create animated GIFs from text or a photo, then let users adjust playback speed and edit individual frames. This feature arrives with Images 2.6, rolling out on web, desktop, and mobile apps, as well as in Codex. OpenAI thus pushes its tool further toward visual editing, not just static image generation.
Du prompt au GIF, sans quitter ChatGPT
Until now, ChatGPT did not offer a native one-step path to a final GIF. Now, a prompt like “can you make a gif of a cat dancing?” can produce an animation, with the option to then adjust speed and inspect frames one by one. The logic changes: it’s no longer just generating an image, but also editing its motion.
This evolution matters for rapid production uses, whether social content, visual mockups, or creative prototypes. OpenAI also notes the new version improves image generation speed throughout ChatGPT. A small but crucial detail: for the chat to become an animator, visual coherence had to hold from frame to frame.
Images 2.6 accélère aussi la cadence
OpenAI presents Images 2.6 as an update focused on speed and flexibility. Beyond GIFs, the company also launches two API models, GPT-Image-2.6 Flare for faster generation and GPT-Image-2.6 Sunburst for higher visual quality. This split follows a logic already seen in Images 2.5, where the choice was between throughput and finer rendering.
The figures give an idea of the product’s pace. OpenAI now claims over 4 billion images per week via ChatGPT and its API, compared to over 3 billion in September 2026 according to previous announcements cited by The Verge and The Decoder. These volumes aggregate all image uses, not just the new GIF function.
Une brique de plus dans la boîte à outils visuelle d’OpenAI
The real interest is strategic. By adding editable animation, OpenAI brings ChatGPT closer to a more complete creation tool, capable of chaining generation, correction, and sharing in one flow. For developers, the API keeps separate options for speed and quality. For users, the result is easier to handle. And less copy-pasting is never lost.
One important limitation remains: sources do not detail supported formats or exact constraints on GIF duration or size. At this stage, OpenAI also does not describe a differentiated rollout by market, with deployment announced for all ChatGPT and Codex users on web, desktop, and mobile.
By the numbers
- 4 billion images per week — volume claimed by OpenAI, November 2026.
- 3 billion images per week — volume announced in September 2026 by OpenAI.
- 2 API models — Flare for speed, Sunburst for quality.
- 3 platforms — web, desktop, and mobile for ChatGPT and Codex.