OpenAI's latest image generator, powered by the GPT-4o model, is now available to anyone using ChatGPT. The announcement was made directly by the company's CEO, Sam Altman, via a post on the X social network, marking a major change of direction compared to the previous limitation to subscribed users only.
Although the precise details on the number of images that users of the free version will be able to create have not yet been disclosed, Altman had mentioned last week a limit of three images per day. It remains to be seen whether this indication will be confirmed when it is opened to all users.
The launch of this image-generation tool was immediately met with enormous success. Altman himself commented that the demand was so high that it put a strain on the company's graphics processing units (GPUs). Quantifying it, it seems that there was a decidedly unexpected influx of new subscribers: we are talking about approximately 1 million new subscribers in just one hour.
Alongside this excitement, the tool has also attracted attention for a particular use: transforming photographs in the style of Studio Ghibli, the famous Japanese animation house. This trend has raised questions about copyright and the data used to train the model, given the marked stylistic similarity.
There has also been no shortage of improper uses, such as creating fake receipts, for example, from restaurants. An OpenAI spokesperson told TechCrunch that all images generated by ChatGPT are marked with metadata indicating their origin and that the company takes action if the images violate its guidelines.
Alongside this important news for users, OpenAI also announced that it has raised $40 billion (about €37 billion) in funding led by SoftBank, reaching a valuation of $300 billion (about €277 billion). The company also announced that ChatGPT has reached 500 million weekly active users and 700 million monthly active users, figures that testify to its growing popularity and diffusion.