OpenAI has launched ChatGPT Images 2.5, introducing a new "Sketch" feature that lets users draw a rough doodle inside ChatGPT and have it transformed into a detailed AI-generated image based on their instructions. The update also allows users to leave comments directly on parts of an image to request specific changes, offering a more precise way to refine outputs without rewriting entire prompts. This matters because it lowers the barrier to producing polished AI imagery, letting users communicate visual ideas through simple sketches rather than relying solely on text descriptions.

The Sketch tool is activated by typing "@Sketch" into the chat box, which opens a drawing window; a Verge reporter tested it by turning a rough mouse-drawn cat doodle into a realistic photo, then used the new comment feature to change the cat's eye colour. OpenAI says the broader Images 2.5 update produces more natural lighting and richer textures, follows multi-turn editing instructions more accurately, and cuts image generation latency by up to 50% compared with the previous Images 2.0 version. The update is available now for ChatGPT, ChatGPT Work and Codex users across desktop, mobile and web.