OpenAI introduces ChatGPT Images 2.5
ChatGPT Images 2.5 brings sharper detail, more precise editing and faster image generation in ChatGPT. Developers get two API models, GPT-Image-2.5 Flare and GPT-Image-2.5 Sunburst.
What it does
ChatGPT Images 2.5 is OpenAI's new image model, with more natural lighting and textures and better preservation of subjects from reference photos. It edits only what you ask, and stays consistent across many rounds of edits.
New ChatGPT features include Sketch (draw a rough guide with @Sketch), templates, comments placed on images, and shared prompts.
Results and limits
OpenAI says latency is up to 50% lower than Images 2.0. Manus reports Flare runs two to four times as fast as GPT-Image-2 in its evaluations. No independent benchmarks are given in the source.
Worth knowing
It's rolling out to all ChatGPT, ChatGPT Work and Codex users. In the API, GPT-Image-2.5 Flare is the default and Sunburst gives more precision with longer generation times. Images carry C2PA metadata and invisible watermarking.
Try it yourself
@Sketch a rough layout of [your shot: subject position, horizon line], then prompt: "Turn this into a [style] storyboard frame of [scene], [lighting]." Follow up with one focused edit: "Change only [element]."