At I/O 2026, Google announced that the Gemini app will now use a new approach to compute limits, adjusting based on the complexity of user prompts. Many users had mentioned feeling restricted by limits that felt too tight.
The new system tracks usage differently. Now, the app refreshes every five hours until the weekly limit is hit. Google explains that a simple text prompt uses far less computing power than a heavy video or coding request. To make it easier for users, Gemini will also offer a pay-as-you-go option for extra AI credits.
Josh Woodward, Gemini’s lead, recently shared that there will be a cap on how much a single prompt can use. This change aims to help users get more out of the Pro model, especially since complex prompts can quickly eat through limits. Google assured users that if an AI request fails, it won’t count against their quota. Successful completions are what matter.
Heavy tasks, like deep research, require more resources. To help users manage their limits better, Google plans to provide more detailed usage reports and notifications. Currently, users can check a basic overview on the Gemini usage dashboard.
In other updates, Free Flash-Lite prompts won’t affect usage quotas anymore. Google will also remember your model choices across sessions, which means it’ll only change if you decide to manually adjust it or hit a limit.
Additionally, Google fixed a bug where just one or two Omni videos could drain resources for some users. Now, Gemini AI Ultra users can generate double the number of Omni videos.
These updates reflect Google’s push to improve user experience and efficiency in AI usage. As technology continues to evolve, staying informed about these changes can help users leverage AI tools to their fullest.
