OpenAI launches Images 2.6, letting ChatGPT generate and edit animated GIFs frame by frame, with GPT-Image-2.6 Flare and Sunburst in the API
OpenAI released its Images 2.6 image model on 4 November, rolling it out to all ChatGPT and Codex users on web, desktop and mobile. According to The Verge, ChatGPT can now turn a prompt or an uploaded photo straight into a GIF, then adjust its playback speed and edit individual frames before sharing it; the API also gets two models, GPT-Image-2.6 Flare and Sunburst.
Until now ChatGPT could not natively create a GIF in one step. In OpenAI’s example, the prompt “can you make a gif of a cat dancing?” produces several slight variations of the same character that are strung together flipbook-style. The Decoder notes this relies on the new model keeping the character consistent across frames, where earlier models would likely have changed its look too much between frames.
OpenAI says Images 2.6 also makes regular image generation in ChatGPT faster. For developers, the API offers two tiers: GPT-Image-2.6 Flare for speed and Sunburst for quality. According to The Decoder, this two-tier split first appeared with Images 2.5.
OpenAI also says users now generate more than 4 billion images a week across ChatGPT and the API, up from the 3 billion a week it reported in September. The openai.com post returned 403 to our direct fetch, so these details rely on reporting by The Verge and The Decoder.