Google has revamped how it measures AI usage for its Gemini service, moving from counting prompts to metering computing power, affecting limits and subscription plan benefits.
- Usage billed by computing power, not request count
- Four subscription tiers offer scaled access and context sizes
- Usage resets every five hours with a weekly cumulative limit
What happened
Google has introduced a fundamental change to how AI usage is tracked for its Gemini platform. Instead of counting the number of prompts users make, Gemini now measures usage based on the computational resources required to fulfill those requests. This means that more complex tasks consume more of a user’s AI quota, potentially lowering the number of responses a user can generate daily depending on the demand of each prompt.
The company also clarified the subscription options available. Alongside the free tier, Google offers AI Plus, AI Pro, and AI Ultra plans with pricing ranging from $8 to $200 monthly. Each tier increases the amount of AI usage available, the sophistication of the AI models accessible, and the maximum context window size—the amount of text handled per interaction. These changes aim to better align usage with operational costs and resource demands.
Why it matters
This shift to metering usage by computational cost rather than prompt count reflects Google's effort to manage capacity and expenses more effectively in supporting AI services. For end users, especially those on free or lower-tier plans, this introduces variability in how many AI requests they can make, as simpler prompts consume fewer resources than complex ones like coding or video generation.
The tiered subscription model that accompanies this change offers scaled benefits but also requires more awareness from users about how their AI usage accumulates. The lack of fixed prompt limits and the possibility of daily fluctuations based on testing and capacity introduce unpredictability, impacting workflows that rely on consistent AI availability. Users must track usage actively to avoid disruptions.
What to watch next
Users should monitor their AI usage via Google Gemini’s app interface, which displays both current usage (resetting every five hours) and weekly totals. These dashboards will become essential in managing how and when users interact with the AI, especially as Google may adjust limits dynamically depending on demand and resource availability.
Future developments may include more transparent communication on usage costs and potentially new subscription options adjusting quota or pricing. Observing how Google balances user experience with operational constraints will be key, especially as competitors in the AI space also evolve their pricing and metering strategies.