The surge in popularity of ChatGPT's new 4o Image Generation feature, launched on March 25, 2025, prompted OpenAI to implement temporary restrictions just two days later12. CEO Sam Altman announced that free tier users would soon be limited to three image generations per day, while the company works on improving the system's efficiency3. These measures are intended to be short-term, with Altman expressing hope that the limitations "won't be long"1. The rapid implementation of these restrictions highlights the unexpected scale of user engagement with the new feature, pushing OpenAI's GPU infrastructure to its limits45.

The viral trend of creating Studio Ghibli-style anime images using ChatGPT's new 4o Image Generation feature has taken the internet by storm. Users have been transforming famous moments, including iconic movie scenes and news photos, into charming anime-style visuals reminiscent of the beloved Japanese animation studio's aesthetic12. This unexpected surge in popularity has contributed significantly to the overwhelming demand that led to OpenAI's GPU infrastructure being pushed to its limits34. The trend showcases the creative potential of AI image generation tools while simultaneously highlighting the challenges faced by AI companies in managing sudden spikes in user engagement and resource utilization.

ChatGPT's new 4o Image Generation feature, powered by the GPT-4o model, offers several advanced capabilities that set it apart from its predecessor, DALL-E 3. The upgraded system can now natively create and modify images, including those containing people, with improved accuracy and detail1. Users can edit existing images, transform them, or 'inpaint' details like foreground and background objects1. While the image generation process takes slightly longer than DALL-E 3, it produces more precise results1.

One notable aspect of the new system is its apparent lack of restrictions on generating images of celebrities, a limitation present in many other AI image generators2. However, OpenAI has implemented "heightened restrictions" around nudity and graphic violence to ensure responsible use of the technology3. The company has also introduced a multi-tiered safety system to limit the generation of violent, hateful, or adult content4, balancing creative freedom with ethical considerations.

Despite its popularity, ChatGPT's new image generation tool faces several technical challenges. Users have reported difficulties in precisely editing specific regions of images, maintaining edits to faces across iterations, and handling tasks such as cropping, data visualizations, and multilingual text rendering12. These limitations have prompted OpenAI to address some issues where appropriate image requests were mistakenly being rejected3. The company has acknowledged these shortcomings and is actively working on improvements, with fixes expected to be implemented in the near future12.