OpenAI has launched ChatGPT Images 2.0, a new version of its image generator that shifts the product closer to a design and visual-workflow tool, not just a prompt-to-picture toy.

What changed

According to OpenAI's launch post and follow-up reporting from The Verge and TechCrunch, the update improves some of the areas that have historically made AI image tools unreliable for practical work: text rendering, instruction following, fine detail preservation, and layout control. OpenAI also says the model now supports up to 2K resolution, a wider range of aspect ratios, and much stronger handling of non-Latin text, including Japanese, Korean, Hindi, and Bengali.

The more notable change is a new thinking-enabled workflow. OpenAI says paid users can let the model reason through a composition before generating it, pull information from the web for current context, and create multiple consistent images from one prompt. The Verge reported that the system can generate up to eight images at once while keeping characters, objects, and styles aligned across scenes.

Why it matters

This is less about prettier AI art and more about whether image generation can become useful for real work like slides, explainers, comics, mockups, and marketing assets. OpenAI says Images 2.0 is rolling out to all ChatGPT and Codex users, with more advanced output options reserved for paid tiers. If those claims hold up in broader use, image models may start to compete more directly with lightweight design software instead of sitting beside it.