OpenAI launched its new image generator for ChatGPT powered by its GPT-4o model on March 25th. The feature quickly gained virality on social media for its ability to transform images into Studio Ghibli’s art style, which led the company to delay the rollout to free users. In a post on X on Tuesday, company CEO Sam Altman announced that the feature is now available to all free users.

At launch, the latest ChatGPT image generator was restricted to paid users. The feature quickly became so popular that even Altman said, “Our GPUs are melting.” Altman added, “Images in chatgpt are way more popular than we expected (and we had pretty high expectations).” On Monday, Altman posted that ChatGPT added one million users in one hour at a point.

Before today’s broad rollout, Altman announced that OpenAI was introducing rate limits to make it more efficient and work with the latest image generator. He added that the free tire will be limited to only three image generations per day “soon.” As part of the update, the GPT-4o model also got better at following detailed instructions, especially prompts containing multiple requests; it also improved its capability to tackle complex technical and coding problems and improved its intuition and creativity.