OpenAI has rolled out an image generation feature in ChatGPT-4o, making it accessible to all users, including those on free, pro, plus, and team accounts. Almost immediately after the launch, users began experimenting with the tool, amazed by its ability to create soft, dreamlike images that bear a striking resemblance to Studio Ghibli’s signature animation style. This discovery led to an explosion of interest across social media, particularly on X (formerly Twitter), where users eagerly shared their AI-generated creations.
The trend gained even more momentum when notable tech figures, including OpenAI CEO Sam Altman and X CEO Elon Musk, joined in on the fun by generating and posting their own AI-created images. As the viral challenge spread, people started using the feature to transform their profile pictures, reimagine famous movie scenes, and recreate iconic historical moments as delicate, anime-style artworks.
User Complaints Over Image Generation Issues
Despite the excitement, many ChatGPT users, particularly those on the free tier, reported being unable to generate images. Some received messages stating that their requests did not comply with OpenAI’s content policy guidelines, even when their prompts were entirely appropriate. Plus, pro, and team users also experienced difficulties, suggesting that the issue extended beyond just free-tier access.
For free-tier users, the issue seemed to stem from OpenAI’s server load management system, which sometimes switches between different AI models when processing requests. Since not all models support image generation, some free users were inadvertently left unable to access the feature. Meanwhile, others were able to generate images without issues, further fueling confusion.
OpenAI Acknowledges and Addresses the Problem
Acknowledging the widespread complaints, Sam Altman addressed the issue publicly, confirming that OpenAI’s content moderation system was mistakenly blocking some image requests that should have been allowed. He reassured users that the company was actively working on a solution, stating:
“We are refusing some generations that should be allowed; we are fixing these as fast as we can.”
As of Thursday, OpenAI said that most of the technical issues preventing image generation should now be resolved, although some inconsistencies may still persist as adjustments continue.
OpenAI to Implement Limits on Image Generation
While the feature’s viral success has been exciting, it has also created an unexpected problem for OpenAI: GPU overload. Sam Altman admitted that the surge in image generation requests was pushing OpenAI’s hardware resources to their limits, stating:
"Super fun seeing people love images in ChatGPT... but our GPUs are melting."
To address this, OpenAI announced that it would soon introduce temporary rate limits on image generation requests to prevent excessive strain on its GPU resources.
-
Free users will be restricted to just three image generations per day.
-
Pro, plus, and team subscribers will also face some limitations, though OpenAI has not yet specified exact numbers.
-
These restrictions will be temporary, with OpenAI working to optimize its models to handle larger-scale demand.
What’s Next for Image Generation in ChatGPT?
As OpenAI works on scaling its infrastructure, the company is expected to fine-tune and expand the feature to offer even more flexibility and creative potential. Future updates may include:
-
Enhanced control over image styles (e.g., photorealistic, oil painting, anime, pixel art, etc.).
-
More efficient GPU usage, potentially allowing OpenAI to increase the daily limit for free-tier users.
-
Integration with video generation tools, allowing users to create short AI-generated animations.
For now, AI-generated art in ChatGPT-4o remains a widely discussed and sought-after feature, and while some users may experience temporary limitations, OpenAI’s continued improvements suggest that the best is yet to come.
