OpenAI Introduces ChatGPT Images 2.5 With New Sketch Feature
OpenAI has introduced ChatGPT Images 2.5, a major update to its image-generation system that brings sharper image quality, faster generation and more precise editing capabilities. The update was announced on September 8, 2026, and is now rolling out across ChatGPT.
One of the biggest additions is a new feature called Sketch, which allows users to draw directly inside ChatGPT and use the rough drawing as a visual guide for generating a finished image. Users can activate the feature by typing @Sketch and then drawing their idea before describing the desired result.
The feature is designed for situations where explaining an idea through text can be difficult. A simple sketch of a room layout, clothing design, object or character can now serve as a starting point for a much more detailed AI-generated image.
ChatGPT Images 2.5 also improves how the system handles reference photographs. OpenAI says the updated model is better at preserving recognizable features of people and animals while changing their surroundings, appearance or visual style.
Editing has received a major upgrade as well. Users can make more targeted changes without unnecessarily altering other parts of an image. The system is also designed to maintain consistency through multiple rounds of editing, allowing users to refine an image without repeatedly rebuilding it from scratch.
OpenAI says image generation can now be up to 50% faster than Images 2.0, while the new model produces more natural lighting, richer textures and improved results for complex visual instructions.
The company has also introduced templates for formats such as posters and merchandise, giving users a structured starting point instead of requiring them to begin with a blank prompt. Users can additionally place comments directly on images to identify specific areas they want changed.
ChatGPT Images 2.5 is being rolled out to ChatGPT, ChatGPT Work and Codex users across desktop, mobile and web. OpenAI says the feature is available across all ChatGPT tiers, although some image capabilities may vary by plan.
The launch also extends to developers through two new API models. GPT-Image-2.5 Flare focuses on faster, high-quality image generation, while GPT-Image-2.5 Sunburst is aimed at workflows requiring greater precision and more detailed creative control.
With Sketch, reference-based creation and more consistent editing, OpenAI is positioning ChatGPT Images 2.5 as a more interactive creative tool—moving image generation beyond simple text prompts toward a workflow where users can draw, mark, edit and refine their ideas directly.
