Rather than announcing a release date for the ChatGPT GPT-5 upgrade this week, which I expected might come in response to DeepSeek’s V3 launch in China, OpenAI went a totally different route. The AI firm unveiled a brand-new AI generation model available directly inside ChatGPT chats.
The tool doesn’t have an easy-to-remember name, as OpenAI refers to it as “4o image generation.” Notably, you don’t have to pick the model when you want to create AI images from the ChatGPT menu. You just tell the AI what sort of images you want, and 4o image generation delivers them.
That ease of use, combined with ChatGPT’s new features and impressive ability to follow instructions, made the new image generator go viral. In just a few hours, ChatGPT users with access to 4o image generation created all sorts of mind-blowing images with simple text prompts.
ChatGPT already had a built-in image generator, the Dall-E platform. But the new one is better by orders of magnitude. For starters, it can generate text correctly. It can follow detailed instructions and let you refine the AI images with additional commands. You can upload real photos and then use the AI to edit them according to your needs. Unfortunately, the new AI image generation tool can also be abused to create deepfakes, and it can steal copyrighted material. All of that certainly helped it go viral, too — and there’s more bad news as a result: ChatGPT Free users will have to wait a long time to be able to generate images with the AI.
OpenAI said on Tuesday that it would roll out 4o image generation to ChatGPT Plus, Pro, Team, and Free users. Enterprise and Edu users would get it later. Notice that the Free tier was included in the initial rollout plans.
I’m a ChatGPT Plus user in the EU, so I didn’t actually expect the feature to be available in Europe at the same time as the rest of the world. But it turns out that OpenAI made the new tool available worldwide on Tuesday. I was able to take advantage of the 4o image generation right away.
However, Sam Altman took to X to say that ChatGPT Free users will have to wait a while before being able to create images with GPT-4o.
“Images in chatgpt are wayyyy more popular than we expected (and we had pretty high expectations),” the OpenAI CEO said. “Rollout to our free tier is unfortunately going to be delayed for awhile.”
It’s unclear what that means. We have no idea what computing load OpenAI can support right now regarding image generation, or how much that costs.
It makes sense for OpenAI to prioritize access to a viral AI feature like GPT-4o image generation to paid users, at least as long as there’s increased demand for it. We, end-users, have to pay for all AI features, especially advanced tools that require continuous, intensive processing.
However, OpenAI never specified limits for the premium tiers that received access to 4o image generation. ChatGPT Pro users can probably generate more images than the Plus tiers, but we have no figures in place, as is the case with ChatGPT features like Deep Research.
As for ChatGPT Free users, they’ll probably get access to the tool as soon as the hype dies down and we get used to all the fakes ChatGPT can now make for you.
