Click to Skip Ad
Closing in...

OpenAI brings GPT-4o image generation to ChatGPT and Sora

Published Mar 25th, 2025 3:03PM EDT
ChatGPT GPT-4o image generation is rolling out now.
Image: OpenAI

If you buy through a BGR link, we may earn an affiliate commission, helping support our expert product labs.

OpenAI launched GPT‑4o image generation this week, giving the company’s flagship AI model the ability to generate precise, photorealistic images and edit uploaded images. This is also the first time that users will be able to generate images directly within ChatGPT—a feature that has been on many wishlists for years.

“We trained our models on the joint distribution of online images and text, learning not just how images relate to language, but how they relate to each other,” OpenAI explains. “Combined with aggressive post-training, the resulting model has surprising visual fluency, capable of generating images that are useful, consistent, and context-aware.”

You can see the GPT-4o image generation capabilities in action in the video below:

OpenAI says that creating and editing images is no different than talking to ChatGPT. Describe what you want to see and include specifics like aspect ratios or hex codes. Because the images are so detailed, they might take up to one minute to render.

Perhaps the most impressive improvement of 4o’s image generation is its ability to render text. One of the telltale signs of an AI-generated image has long been garbled, nonsensical text. GPT-4o is smart enough to know how to not only render English words but also put them in the right order. You can see one impressive example below:

4o image generation can also render text accurately.
4o image generation can also render text accurately. Image source: OpenAI

4o’s image generation is also capable of building upon images and text in chat context, following detailed prompts with attention to detail, analyzing and learning from user-uploaded images, and linking its world knowledge between text and images.

Of course, it’s not perfect. Some of the image generator’s issues include cropping long images too tightly, making up information, and struggling to render non-Latin languages.

4o image generation is rolling out now for Plus, Pro, Team, and Free users as the default image generator in ChatGPT. Enterprise and Edu will gain access soon, and developers will be able to generate images with GPT‑4o via the API in the coming weeks. It’s also available in Sora and even through a dedicated DALL·E GPT for DALL·E diehards.

Jacob Siegal
Jacob Siegal Associate Editor

Jacob Siegal is Associate Editor at BGR, having joined the news team in 2013. He has over a decade of professional writing and editing experience, and helps to lead our technology and entertainment product launch and movie release coverage.