12:00
10:00
09:00
20:30
16:00
11:00
12:00
10:00
09:00
20:30
16:00
11:00
12:00
10:00
09:00
20:30
16:00
11:00
12:00
10:00
09:00
20:30
16:00
11:00
OpenAI has released ChatGPT Images 2.5, a new image-generation model that can create images up to twice as fast as its predecessor.
According to OpenAI, generation latency has been cut by 50% compared with Images 2.0, while the new model is also better at preserving a person’s appearance and object details from reference images, reproducing textures more accurately and maintaining more natural lighting.

The model is also more reliable when making targeted edits, so changing a single element such as a background or a piece of text is less likely to affect the rest of the composition. In longer editing sessions, it can retain context from earlier changes without losing image quality across iterations, and it now supports more complex compositions with transparent backgrounds.
OpenAI is updating several image tools in ChatGPT alongside the new model:
@Sketch, which the model can then use as the basis for a finished image.Another small update is that image generation now shows a progress percentage while the result is being created.

For developers, OpenAI has released two API variants: GPT-Image-2.5 Flare, focused on faster generation, and GPT-Image-2.5 Sunburst, designed for higher-detail output.
ChatGPT Images 2.5 is now available across ChatGPT, ChatGPT Work and Codex on desktop, mobile and the web. OpenAI recently also rolled out its latest flagship language model, GPT-6 Astra.


