OpenAI announced Images 2.6 on November 4, 2026, saying the update brings GIF creation to ChatGPT alongside changes to image generation in Codex and its API. The company describes a feature that turns written prompts or uploaded still images into looping animations.
OpenAI said Images 2.6 was available globally across ChatGPT and Codex subscription tiers on desktop, mobile and web. Its help documentation describes GIF creation as rolling out in ChatGPT.
GIF creation from prompts, images and sketches
OpenAI says users can describe the movement they want in a looping GIF or upload a still image and ask ChatGPT to animate it. The company also demonstrated a hand-drawn owl sketch becoming an animated character, illustrating how a rough drawing can serve as a starting point.
The feature extends ChatGPT image generation into short animations. OpenAI lists animated storyboards for exploring character movement before video production among its intended uses. It also describes applications involving reaction images, brand mascots and product visuals.
For GIFs created in ChatGPT, OpenAI says users can request changes in the conversation and open the animation to adjust playback speed, save it or share it. The company’s help documentation states that ChatGPT cannot edit an animated GIF uploaded by a user.
OpenAI also documents an option to save an uploaded photo as a named reference. Users can refer to that image in a later conversation by typing the @ symbol followed by its saved name. This provides a way to reuse a visual reference across image requests.
Speed claims and API availability
OpenAI claims Images 2.6 generates images 50 percent faster than its predecessor, Images 2.5. It says GPT-Image-2.6 Flare, the version focused on speed, can generate an image in as little as six seconds when quality is set to low. The six-second figure applies to low-quality image generation.
The company’s help documentation says image generation can take a few minutes, depending on the complexity of the request.
OpenAI said GPT-Image-2.6 Flare and GPT-Image-2.6 Sunburst were available through its image API. The company positions Sunburst for work requiring greater detail and consistency across edits at medium or high quality.
OpenAI’s API documentation describes animated GIF output for both versions, including generation from text and animation guided by still reference images. Developers can specify frame count and timing. The same update adds controls for how much planning goes into an image, markup-guided editing and SVG output, according to OpenAI.
