Google has begun rolling out a new usage restriction system for Gemini that fundamentally changes how limits are calculated. Instead of simply counting the number of messages or queries a user sends, the new model takes into account the complexity of each request, which AI model is being used, and the amount of computing resources the task consumes.
In practice this means that two users who send the exact same number of messages could use up very different portions of their allowance. A straightforward question consumes relatively little, while processing large documents, running Deep Research, generating images or video, or handling complex coding tasks will drain the quota far more quickly.
Limits Reset Every Five Hours
One of the most significant changes concerns how usage resets. Rather than a standard daily limit, Google is introducing five-hour cycles during which a portion of the resources is renewed. However, users will also be subject to a weekly cap, meaning those who use Gemini intensively could find themselves locked out of certain features until the next billing period begins — a frustration that has already prompted complaints from power users.
Google’s position is that this model is fairer, since it measures actual load on the infrastructure rather than a blunt count of messages sent.
A New “Usage Limits” Dashboard
To help users navigate the new system, Google is adding a dedicated section inside Gemini called “Usage Limits.” From there, users will be able to monitor how much of their quota they have consumed, when the next renewal is due, the status of their weekly cap, and how many resources they are consuming in real time. This marks the first time Google has given users this level of visibility into their consumption, rather than simply blocking access once a limit is hit without warning.
Power Users Are the Most Affected
The changes have landed hardest on users who rely on Gemini for professional work — developers, researchers, and creators producing multimedia content. Several have reported that even paid subscription tiers are becoming too restrictive, with weekly quotas potentially exhausted after just a few intensive work sessions.
The change has started, as Google are making adjustments to some of the limits based on user reactions, including increasing the quotas on their AI coding tool Antigravity after user complaints. It remains to be seen whether further adjustments will follow for other features.
A Tiered Future for AI
The new rules draw a clearer line between different subscription levels. This tiered structure is that the free user gets one amount, AI Plus gets two, AI Pro gets four, and then there is AI Ultra designed for those that really need a large amount of the system. This does more than just affect Gemini, though. It reflects a trend taking hold across the entire AI industry: the era of seemingly unlimited access is quietly giving way to a model based on credits, quotas, and consumption-based billing. The infrastructure costs of running large AI models are enormous, and companies are increasingly looking for ways to make those costs visible to users rather than absorbing them behind a flat-rate subscription.
Google argues this approach distributes resources more equitably. Critics, however, point out that it turns what was once a simple tool into something that requires users to constantly monitor their consumption — calculating the cost of every query before sending it, rather than simply working. Whether that trade-off is worth it will likely depend on which tier you’re on and how heavily you rely on AI in your daily workflow.
