At I/O 2026 last week, the Gemini app changed to computing-based usage limits. In answer For “comments about reaching limits too quickly,” Google announced some changes today.
The new “computer use” usage approach (5 hour refresh until weekly limit is reached) aims to take into account the complexity of the prompts, the tools used and the duration of the chat. Last week, Google noted that “a simple text message uses much less computing than a complex video or encrypted message.” In the future, Google will allow Gemini app users to purchase refill AI credits per use.
Using Gemini 3.1 Pro, Gemini leads Josh Woodward today shared that Google is “limiting the amount of quota a single message can use to get more out of the Pro model.” This is in response to complex prompts with large files that quickly exhaust limits.
Google clarified that errors do not count toward limits: “If a request fails, you will not be charged. Errors in our system are our fault, not yours. Your quota is used only for successful completion.”
Heavy tasks like Deep Research “require more compute,” so Google will provide “more detailed usage breakdowns and notifications to help you maximize your limits.” As it is, the gemini.google.com/usage The dashboard only provides a high-level overview.
Meanwhile, Flash-Lite 3.1 prompts are now “free and will not count toward your quota.” Google also points out how:
When you select a specific model, we remember that choice in all future sessions. It will only change if you adjust it manually or press a limit that triggers an automatic downgrade to a lighter model.
Finally, Google fixed a bug where “just one or two Omni videos” would exhaust “certain people’s” quotas. Google AI Ultra users have now doubled the number of Omni generations.
We fixed this issue and will continue to look for opportunities to increase the amount of Omni you earn.
FTC: We use automatic affiliate links that generate income. Further.
