Image: OpenAI
OpenAI has launched Images 2.6, an update to image generation in ChatGPT, Codex and its API that adds animated GIFs and, according to the company, generates images 50% faster than before. In the API, the update arrives as two models called GPT Image 2.6 Flare and GPT Image 2.6 Sunburst.
In its announcement, OpenAI said users can now bring motion to images with GIFs, from animating a brand mascot or character to turning a product shot into a loop. The company pitched the format at marketers and creators for campaign mascots, social media reactions and memes, and animated storyboards that explore character movement before video production. OpenAI also said a simple sketch can become an animated character. Its post on X highlighted reference images as the other headline addition.
Images, upgraded—introducing ChatGPT Images 2.6.
— OpenAI (@OpenAI) November 4, 2026
Now with GIFs and reference images.
Plus, two new models are coming to the API with 50% faster image generation. pic.twitter.com/kcuOhTipYR
OpenAI describes Flare as the option for speed. When quality is set to low, the company says Flare can generate an image in as little as six seconds, and it also claims better layout quality for posters and product imagery. Sunburst is aimed at work that needs more detail, with OpenAI saying it offers higher quality output and stronger consistency across edits at medium or high quality settings.
Developers also get several new tools. Image markup, already available in ChatGPT, is coming to the API so teams can mark exactly where an image should change. OpenAI's image generation guide says markup takes a transparent PNG overlay matching the size of the first input image. The API also adds SVG export and a new complexity setting that controls how much planning the model does before rendering, with low suited to quick edits. Both models support these features.
According to the guide, GIFs are requested through the generations endpoint and can run from 2 to 64 frames, with each frame shown for 20 to 2,000 milliseconds. GIF output is not supported on the edit endpoint. On the Sunburst model page, OpenAI lists image output at $30 per million tokens and image input at $8 per million tokens.
OpenAI also shared usage figures, saying people now create 4 billion images per week across ChatGPT, Codex and the API, which it put at more than half a billion a day. Early partners quoted in the announcement include OpenArt chief executive Coco Mao and Figma Weave creative director Itay Schiff, who said the model will be added to Figma Weave.
The launch follows signs of the model that surfaced last week, when GPT Image 2.6 appeared in OpenAI's API documentation in the description of the markup option before any announcement.
OpenAI says Images 2.6 is available globally in ChatGPT and Codex across all subscription tiers on desktop, mobile and web, and both models are live in the API at the time of this writing. The company has published a system card covering its safety evaluations for the release.
Comments