ChatGPT Now Supports Image Generation With GPT-4o

OpenAI is integrating image generation capabilities right into ChatGPT, and it’s doing so using its natively multimodal GPT-4o model. The company said that ChatGPT users now have access to its “most advanced image generator” yet, with GPT-4o being enabled by default for Plus, Pro, Team, and Free users.

You may remember OpenAI’s DALL·E image generator, which is still available through a dedicated DALL·E GPT, but GPT-4o image generation in ChatGPT should deliver much better results. Users can also refine the images they want to create in ChatGPT with a series of prompts.

“GPT‑4o image generation excels at accurately rendering text, precisely following prompts, and leveraging 4o’s inherent knowledge base and chat context—including transforming uploaded images or using them as visual inspiration,” OpenAI explained today. “These capabilities make it easier to create exactly the image you envision, helping you communicate more effectively through visuals and advancing image generation into a practical tool with precision and power.

OpenAI said that GPT-4o can accurately create up to 20 different objects in an image while “other systems struggle with ~5-8 objects.” However, the company also acknowledged that GPT-4o currently has some limitations, including hallucinations with low-context prompts and text rendering issues in images with dense information and small text. The company is already planning to address these limitations through post-launch model improvements.

While image generation in ChatGPT will roll out to free users with usage limits, it will also be coming soon to Enterprise and Education users. Developers will also be able to use OpenAI’s API to generate images with GPT-4o in the coming weeks.