OpenAI on September 8 launched ChatGPT Images 2.5, saying the update brings sharper detail, more reliable multi-turn editing, better subject preservation from reference photos, and up to 50% lower image-generation latency than Images 2.0. The company said people create more than 3 billion images a week across ChatGPT Images and GPT-Image models in the API. Images 2.5 is rolling out to ChatGPT, ChatGPT Work, and Codex on desktop, mobile, and web, per the OpenAI launch post.
The Verge reported a new Sketch tool: typing @Sketch opens a canvas so users can doodle a composition and then prompt from that sketch, plus inline comments on image regions for targeted edits. OpenAI also highlighted templates and prompt sharing. For developers, the API adds GPT-Image-2.5 Flare as the default faster model and GPT-Image-2.5 Sunburst for higher-precision, slower creative workflows.
This brief covers the September 8 OpenAI launch post and The Verge’s Sketch report; it does not re-score image quality claims beyond the company’s stated latency and editing improvements.