AI Tools Are Melting GPUs: OpenAI and Google Slash Daily Limits on Sora and Nano Banana Pro
The global rush toward AI-generated videos and images has pushed demand to unprecedented levels. As a result, both OpenAI and Google have now introduced strict daily usage limits for their flagship media-generation tools—OpenAI’s Sora video generator and Google’s Nano Banana Pro image model.
OpenAI Cuts Free Sora Video Generations to 6 Per Day
OpenAI has officially restricted free users of the Sora app to just six video generations per day. These limits, however, do not affect paid ChatGPT Plus or Pro users, who continue to enjoy higher or unchanged generation capacities.
Bill Peebles, OpenAI’s Head of Sora, confirmed the change on X, saying: “We’re setting usage limits for free users to 6 gens/day. chatgpt plus and pro users have unchanged limits, and everybody can purchase additional gens as needed.”
He also acknowledged the overwhelming surge in demand, humorously adding that “our GPUs are melting, and we want to let as many people access Sora as possible!”
Google Restricts Nano Banana Pro to 2 Image Generations for Free Users
Google is implementing similar restrictions for its rapidly viral image model, Nano Banana Pro, which is powered by Gemini 3 Pro Image technology. Free users are now limited to two image generations or edits per day unless they subscribe to a Google AI premium plan.
In addition, Google appears to have introduced new, variable usage tiers for Gemini 3 Pro. According to a support page referenced by 9to5Google, free access now falls under “Basic access – daily limits may change frequently,” indicating dynamic throttle controls based on traffic.
Older Nano Banana Model Still Allows 100 Images Daily
Interestingly, the previous generation Nano Banana—powered by Gemini 2.5—still permits up to 100 image generations per day. This gives users an alternative if they want lighter AI workflows without hitting the new Pro limitations.
As competition intensifies and GPU demand skyrockets, it is likely that more AI platforms will begin to enforce similar usage controls to balance server load, maintain output quality, and ensure fair access for all users.
