OpenAI's ChatGPT image generator now creates more detailed and consistent pictures — and people are testing its limits with weird requests

OpenAI released ChatGPT Images 2.0, an upgraded image generation feature that produces higher-quality images with better detail, more natural hands and faces, and stronger consistency when you ask for multiple related images in one prompt.

The improvement matters because creators, designers, and business teams can now use AI-generated images for real work instead of just drafts — marketing materials, concept art, and product mockups require fewer human touch-ups, saving hours per project.

You can start using it in ChatGPT today if you have a paid account (Plus or Pro subscription); free users still get limited access. The update also handles complex, specific requests better—like 'a raccoon operating a ham radio' — which previously would have produced blurry or nonsensical results.

The leap in capability means stock photography and junior design roles face new pressure: clients can now generate custom images instantly rather than hiring photographers or designers for routine work, which will reshape hiring in creative industries over the next 12–24 months.