OpenAI began rolling out Images 2.6 on 4 and 5 November, adding GIF creation to ChatGPT and Codex on web, desktop and mobile. A GIF can start from a text prompt or an uploaded photo, and users can change playback speed and step through frames to edit. Suggested uses include mascots, product shots, sketches, reactions and storyboards turned into short loops.
Developers get two new API models. GPT-Image-2.6 Flare is the speed-and-layout option for posters and product images, finishing in as little as six seconds at low quality. GPT-Image-2.6 Sunburst favours detail, and OpenAI says it keeps edits more consistent from one to the next at medium or high quality.
The ChatGPT logo on a smartphone screen, the article's header picture
Shutterstock/Decrypt · Image: Decrypt
The API also gains ChatGPT's image markup, so teams can mark the exact area to change, along with SVG export and a complexity setting. Reference images are the other addition: pictures of yourself, pets or characters can be saved once and reused. OpenAI says generation is 50% faster and cites 4 billion images a week.
About Images 2.6
Images 2.6 follows Images 2.5, which OpenAI released in September 2026 with sharper detail, more precise editing and generation up to 50% faster.
In Decrypt's earlier comparison, Images 2.5 and Google's Nano Banana 2 each won three of six test categories. Images 2.5 was strong on illustration but still stumbled over text.
Nov 2026
Released
Availability is uneven. TestingCatalog describes a global rollout across every tier, yet Decrypt found most users still without GIFs and no date for all accounts. Decrypt, citing The Verge, also noted no standalone announcement. It follows Images 2.5 from September and the March shutdown of the Sora app; GIFs are far simpler than video, but they bring motion into ChatGPT again. Decrypt added that it could not test the model's performance itself.
Sources · 2