"ChatGPT Images 2.0" Announced: An Image Generation Model Where AI 'Thinks Before It Draws,' with More Accurate Japanese Text

OpenAI unveiled ChatGPT Images 2.0, a new image model that can review composition and accuracy before generating.

The model improves rendering for non-Latin scripts, including Japanese, making text-heavy images more practical.

It can generate up to eight images at once and is now available via API.

The update is expected to boost real-world use for posters, charts, comics, and other text-based visuals.