OpenAI rolls out ChatGPT Images 2.5, cuts lag up to 50%

OpenAI has released an updated version of its image-generation tool, ChatGPT Images 2.5, which boasts faster generation times, sharper details, and more refined edits. The new tool enhances the natural lighting and textures in generated images and improves its ability to follow user instructions across multiple turns. It also introduces a new tool called Sketch, which allows users to sketch out images using their fingers on a touchscreen as a reference point for generation.

ChatGPT Images 2.5 is also better at preserving the details of subjects in reference photos and maintains the overall stability of the image during iterative edits. The update includes two variants: GPT-Image-2.5 Flare, which prioritizes speed and creates high-quality images in a fraction of the time, and GPT-Image-2.5 Sunburst, which prioritizes precision and generates images with greater detail.

OpenAI has also implemented security measures, such as invisible watermarking and C2PA metadata, to prevent unauthorized reuse and deepfakes. The new model is currently available across various platforms, including desktop, mobile, and web versions of ChatGPT, ChatGPT Work, and Codex.