OpenAI’s new image generator, powered by its GPT-4o model, is now available to all users, CEO Sam Altman said in a post on X. The feature was until now available only to paying users of ChatGPT.

While it is not clear how many images users on the free tier can generate, Altman last week had mentioned a limit of three images per day.

OpenAI’s image generation tool took off instantly after launch, with Altman saying the demand was so high, the company’s GPUs were “melting.” The tool quickly also gained notoriety for being used to convert pictures into the style of Japanese animation firm Studio Ghibli, raising concerns around copyright and training data used by the company as well, given the similarity in style.

Some people also used it to generate fake receipts, such as restaurant bills. An OpenAI spokesperson told TechCrunch that all these images have metadata indicating that ChatGPT generated them, and that the company “takes actions” if the images violate the company’s guidelines.

Meanwhile, OpenAI today said it raised $40 billion in funding led by SoftBank at a $300 billion valuation. The company also said ChatGPT has hit 500 million weekly active users and 700 million monthly active users.