On 2026-09-10, OpenAI released the image generation model ChatGPT Images 2.5, as well as GPT-Image-2.5 Flare and GPT-Image-2.5 Sunburst for API users.

ChatGPT Images 2.5 improves the reproduction of natural light and textures, and introduces a feature that allows users to develop new styles and compositions while maintaining the subject of a reference photo. It also supports precise editing, allowing specific areas to be modified while preserving surrounding details. Furthermore, generation latency has been reduced by up to 50% compared to the previous Images 2.0.

As production support features, "Sketch," which allows direct drawing within ChatGPT, and templates that enable starting from formats such as posters have been added. These additions facilitate image creation using rough sketches as visual guides and make the personalization of information easier.

For the API, OpenAI provides "Flare," which enhances quality and speed, and "Sunburst," which enables strict control over editing units. Regarding safety, the company continues to employ C2PA metadata and invisible watermarking in addition to prompt and image-based checks.

Source: