ChatGPT generates animated GIFs from an image or sketch

OpenAI adds looped GIF creation to Images 2.6, its image generator integrated into ChatGPT. The feature is rolling out gradually, with generation announced as 50% faster.

On November 4, 2026, OpenAI announced Images 2.6, a new version of its image generator integrated into ChatGPT and Codex that adds looped GIF creation. The tool can produce an animation from a description, an image, or a sketch, then edit it within the conversation before download. This evolution targets social uses, meme creation, and quick-share content, where a static image is no longer enough.

Des GIF directement dans la conversation

The most visible novelty is simple: asking ChatGPT to animate an image. OpenAI states it is enough to specify the desired movement, such as "Create a GIF of a pink cat dancing." From an imported image, the model first proposes a version with several poses, then generates the final animation.

The service mainly targets reaction uses, animated mascots, or small social content. In practice, this brings visual generation even closer to a flexible editor, where you adjust rather than restart. OpenAI also says it accepts a simple sketch sent in the conversation. The GIF finally makes its entrance in meetings.

Retouches, téléchargement et déploiement progressif

Once the animation is created, the user can request modifications in the chat before downloading it. OpenAI also specifies that the feature is being rolled out: it may not be immediately visible on all accounts. The company announces a global launch on all ChatGPT and Codex plans, but without instant availability for everyone.

The Images 2.6 version should also speed up image generation by 50% according to OpenAI, while better preserving details during edits. The company adds a handy feature for recurring uses: saving a reference image under a name, then recalling it in another conversation with a mention preceded by @. This avoids resending the same photo with each new request, which, over time, would have tired any digital dog.

Ce que change Images 2.6 pour les usages courants

By combining generation, animation, and editing in the same thread, OpenAI aims to simplify a use that previously required several tools. For marketing teams, content creators, or internet users tinkering with visuals for social networks, the gain is mainly in time and reduced friction. The usual question remains: ease of use alone says nothing about the final result’s quality.

At this stage, examples shown by OpenAI mainly cover short animations and simple loops. The tool accepts static images, descriptions, and even sketches, but the rendering may vary depending on the source. The promise is clear: to make visual creation faster and more flexible. Deployment will show whether the feature becomes a reflex or just another gadget in chatbot toolkits.

To read