OpenAI’s updated image generation tools in ChatGPT now allow users to incorporate hand-drawn sketches into their prompts.
OpenAI announced ChatGPT Images 2.5 on Tuesday, introducing a new method to instruct the AI: drawing a doodle directly within the interface. With the new "Sketch" feature, users can draw an object and then describe how they want it transformed into an image.
Accessing this feature is simple; users can type "@Sketch" in the chat box to open a drawing window. During testing, the tool successfully rendered a rough mouse-drawn cat sketch into a realistic photograph based on specific instructions. Users can also leave direct comments on image elements to request changes, such as altering the color of the cat's eyes.
Regarding the underlying technology, ChatGPT Images 2.5 offers improvements including "more natural lighting and richer textures," enhanced ability to follow editing instructions over multiple turns, and a "up to 50% reduction in image generation latency" compared to the previous version. The update is currently available to users of ChatGPT, ChatGPT Work, and Codex across desktop, mobile, and web platforms.
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.