On Tuesday, OpenAI made its image-generation feature, powered by the GPT-4o model, available to all users. This capability was initially offered exclusively to paid subscribers of ChatGPT last month, but now users on the free tier can also take advantage of it. Free users will experience some limitations, such as a reduced usage allowance and higher rate limits. The San Francisco-based company noted that every image generated will include metadata from the Coalition for Content Provenance and Authenticity (C2PA) to address deepfake concerns.
OpenAI Expands GPT-4o Image Generation to Free Users
OpenAI CEO Sam Altman announced the expansion via a post on X (formerly Twitter), indicating that the image-generation feature is now accessible to free-tier users of ChatGPT. It is important for users to be logged into their OpenAI accounts to utilize this feature. Currently, free-tier users are limited to generating three images per day, with a cooldown period enforced between consecutive image requests.
In another update, Altman revealed that one million new users registered for ChatGPT within just one hour on Monday, a remarkable increase compared to the platform’s previous record of one million users in five days following its launch in 2022.
The surge in demand for the new image-generation feature can be attributed to a popular trend where users are utilizing ChatGPT to transform their images into the artistic style of Studio Ghibli, the acclaimed Japanese animation studio known for films like Spirited Away, Princess Mononoke, and My Neighbour Totoro. Numerous posts featuring these “Ghibli-style” transformations have inundated social media platforms.
The overwhelming popularity of this trend prompted Altman to implement temporary rate limits for paid ChatGPT subscribers when generating images. He humorously remarked that the decision was due to “our GPUs are melting.”
In a related context, Altman criticized Meta’s recent restrictions on the use of its open-source Llama models for companies with over 700 million monthly active users. He stated in a post, “We will not do anything silly like saying that you can’t use our open model if your service has more than 700 million monthly active users.”
The new image-generating capabilities in ChatGPT also feature enhanced text rendering alongside a prompt-based inline editing function. Users can upload their own images and request AI-assisted transformations into various artistic styles or edits. OpenAI has emphasized that the model blocks requests aimed at generating harmful content, including child sexual abuse material, sexual deepfakes, and violent imagery.
