GPT Image 2

Up to 9 reference images can be combined in GPT Image 2 to keep characters, styles or product details consistent from one render to the next.

GPT Image 2 is an AI image generator and photo editor that turns text prompts and uploaded references into 4K photorealistic visuals.

It also supports image-to-image editing, where you describe the change in plain language instead of masking or painting it by hand.

Text inside images is a focus: the tool claims 95%+ accuracy for in-image text across 12 languages, which matters for posters and mockups where a misspelled headline ruins the result.

You can set color palettes with HEX codes and adjust lighting and facial features.

Typical jobs include marketing posters, UI mockups, product photography and character design.

The site also lists video generation among its categories, though image work is clearly the core.