OpenAI announced that Images 2.5 reduces generation latency by up to 50% compared to its predecessor, Images 2.0. It also better preserves the appearance of reference objects such as people and pets, and improves consistency in multi-round editing. New features include the ability for users to directly add comments on images and specify modification areas, as well as a Sketch function that allows users to draw a sketch for the AI to generate a complete image. Additionally, OpenAI has released two API models for developers: GPT-Image-2.5 Flare and GPT-Image-2.5 Sunburst.