How Google’s New Gemini Rates Work and How to Track Your Usage (www.wired.com)

Google has recently upgraded its Gemini AI applications, enhancing their capabilities across various services, but it has also implemented a new usage metric that measures AI requests by computing power instead of the number of interactions. This shift means users might find themselves using fewer requests overall, especially if their tasks are complex. As Google explains, this new system aims to better reflect the costs of resource use, although it leaves users uncertain about their limits, given that usage is now contingent on the complexity of prompts and the AI model selected. The Gemini AI service offers four subscription tiers—Free, Plus, Pro, and Ultra—each with varying levels of access and limits based on payment. The specifications are somewhat vague, with free-tier limits described as "standard," and paid plans providing multipliers for usage. Notably, as users progress through the different AI models and their respective context window sizes—from 32K tokens for free users to a million tokens for Ultra users—they unlock higher capacities for conversation threads. Checking usage status has been simplified, but Google warns that limits may change without notice, especially under capacity constraints, highlighting a more dynamic but potentially confusing landscape for AI users.