ChatGPT Images 2.5 is Here With 50% Faster Image Generation and New Sketch Tool

OpenAI has just dropped ChatGPT Images 2.5, and this update brings a serious upgrade to image generation inside ChatGPT. The new model promises sharper visuals, faster generation, more accurate editing, and a new Sketch feature for turning rough ideas into finished images.

ChatGPT Images 2.5 brings faster and smarter image creation

OpenAI says Images 2.5 delivers richer textures, more natural lighting, and sharper details across generated images. The model is also better at preserving recognizable subjects when users provide reference photos.

Generation speed has received a noticeable boost, with OpenAI claiming up to 50% lower latency compared with Images 2.0. That should make creating and refining multiple images considerably quicker.

Editing is another major focus, as Images 2.5 can follow precise instructions while keeping other parts of an image unchanged. OpenAI says the model also maintains better consistency across multiple editing turns.

Sketch lets users draw their ideas directly

The new Sketch feature gives users another way to guide image generation without relying entirely on text prompts. You can draw a rough layout, outfit, room, or simple idea, then ask ChatGPT to turn it into a complete image.

OpenAI is also introducing templates for formats including posters, merchandise, and product photos. Users can now add comments directly to images, making specific editing requests easier to communicate.

Developers are getting two new API models alongside the ChatGPT upgrade. GPT-Image-2.5 Flare focuses on faster generation, while Sunburst targets detailed creative workflows requiring tighter editing control.

ChatGPT Images 2.5 is currently rolling out across desktop, mobile, and web for ChatGPT, ChatGPT Work, and Codex users.