Introducing ChatGPT Images 2.5
Introducing ChatGPT Images 2.5 OpenAI's image generation models are apparently used "more than 3 billion images across ChatGPT Images and the GPT‑Image models in the API". This latest release improves their instruction-following ability across multiple turns, responds faster, and "is better at preserving the subjects in your reference photos". There are two new model IDs in the API:…
OpenAI has unveiled the latest iteration of its image generation models, ChatGPT Images 2.5. The new release boasts enhanced instruction-following capabilities, allowing for multiple turns of interaction, faster response times, and improved preservation of subjects from reference photos. Two new model IDs have been introduced to the API: gpt-image-2.5-sunburst and gpt-image-2.5-flare.
Sunburst is reportedly the superior choice for workflows that require editing precision, while Flare is best suited for fast, high-quality everyday image generation.
The upgrades were tested using a CLI tool, openai_image.py, which now supports the inclusion of reference images. This enables the generation of images based on a prompt and one or more reference photos. For instance, a raccoon scientist studying a chart was successfully added to an original image through this method.
Simon Willison, a tech blogger, shared his experience with the new models on September 8th, 2026. He offered a way for readers to support his work, suggesting a $10 monthly contribution for a curated email digest of the most important developments in large language models (LLMs).
Written by urgent.news from Simon Willison's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
