🤖 ChatGPT-4o can now generate crazy-ass images natively (without DALL-E), which means

🐒 Cherry-pick warning: see "Best of N" at the bottom of each generation

✦ Precise instruction following
✦ Learning styles on the go, e.g. from a bunch of example images
✦ Generating from any input, e.g. read Three.js 3D model code, reason and generate a close sample of how it would look like
✦ Amazing text rendering
✦ Refining because it can see full dialog (including previous generations), meaning it won't produce absolutely different image and can do rather precise changes

Limitations mention poor handling of long images (a lot of numbered entries, long images cropping, etc), graph precision issues, multilingual or small text rendering

4o image generation rolls out starting today to Plus, Pro, Team, and Free users as the default image generator in ChatGPT, with access coming soon to Enterprise and Edu. It’s also available to use in Sora. For those who hold a special place in their hearts for DALL·E, it can still be accessed through a dedicated DALL·E GPT.

Developers will soon be able to generate images with GPT‑4o via the API, with access rolling out in the next few weeks.

➡️ OpenAI