Available starting 25/03/2025 to Plus, Pro, Team, and Free users as the default image generator in ChatGPT.
A bit slowish in my opinion but impressive quality, reality simulation and decent level of prompt intelligence and understanding.
My impressions and take aways are
- very realistic generation
- takes while to generate & render
- very good with text (based upon limited testing and video) FINALLY
- smart, not really complex or long prompts to create good pictures
- works relatively well with uploaded photos
- multi-turn generation: correct or refine images through continued conversations
- API access to be rolled out in the next few weeks.
In short, a big step up from the Dall-E model.
Read more
- Introducing 4o Image Generation for some good prompt examples.
- OpenAI’s release video: 4o Image Generation in ChatGPT and Sora (YouTube)