Washington, May 19 (QNA) - Google has announced a significant and unprecedented change to the way it calculates usage limits for its artificial intelligence (AI) platform, Gemini. The company is moving away from the traditional system based on a fixed number of daily questions and officially adopting an intelligent computing model that utilizes actual computing power (API) consumed.

According to the company, this revolutionary change comes to address the enormous operational costs and technical pressure resulting from operating advanced features and the increasing complexity of requests, such as Deep Research tools, generating images and videos, processing complex code, and preserving the context of long conversations that include millions of words.

Under the new system, simple questions will not be equal to heavy tasks. Instead, the capacity allocated to the user will be dynamically cut based on the weight and sophistication of the model used and the length of the data context. Subscribers to paid plans will receive exceptional computing capacity, up to twenty times that of the free plan, with an accurate indicator that allows users to monitor the density of computing consumed and the periods of automatic renewal of quotas every few hours, reflecting the new trend of major technology companies worldwide to control the cost of supercomputing power and generate sustainable economic value from AI services.

The company said the compute-based usage limit system tracks resource consumption over rolling five-hour windows and an overall weekly quota.

Previously, Google relied on a fixed number of daily requests, as Google AI Pro subscribers received up to 100 requests per day via the Gemini 3.1 Pro model, regardless of the complexity of the requests or the amount of resources required to process them.

These changes reflect the challenges facing major AI companies as the capabilities of Agentic AI systems grow, which are able to create sub-agents that consume tens of thousands of code snippets during a single conversation. (QNA)