Earlier this summer, Google introduced a wave of upgrades to its Gemini AI applications, significantly expanding their capabilities. The company's artificial intelligence now spans more of its services, making it increasingly difficult to interact with any Google product without encountering an AI feature or assistant. Alongside these enhancements, however, Google has also imposed new restrictions on how much users can do, completely revamping the way AI usage is calculated and distributed across its subscription tiers.
How AI Usage Is Now Measured
The most significant change is that Google no longer counts the number of requests a user makes. Instead, the company measures usage based on the computing power required to fulfill each prompt. This means that a user who previously could generate three videos per day might now only manage two, particularly if those videos are complex and demand more processing resources.
From Google's perspective, this approach is more logical because it tracks the actual cost to the company's data centers. For end users, however, the system is considerably less transparent. People can no longer rely on straightforward rules such as a fixed number of image generations per day, and it becomes harder to predict when a limit will be reached.
Google's own support documentation adds another layer of uncertainty. The company states that access may change or be limited based on testing, experimentation, or availability, meaning that the amount of AI prompting a user can do may vary from one day to the next. Several factors influence how much AI a person can use: the subscription plan they are on, the length and complexity of their prompts, and the specific Gemini model they select. A simple weather forecast request, for instance, consumes far less computing power than asking the AI to write a complete mini-application.
Subscription Tiers and Their Quotas
In the United States, users can choose from four Gemini AI subscription options or remain on the free tier. The paid plans are AI Plus at $8 per month, AI Pro at $20 per month, and AI Ultra, which costs either $100 or $200 per month depending on the level selected. As users pay more, they gain access to more advanced AI models for longer periods.
Google does not publish specific numerical limits for the free tier, describing them only as standard. AI Plus subscribers receive twice those standard limits, while AI Pro subscribers get four times the standard amount. AI Ultra limits are either five times or twenty times higher than those on AI Pro, depending on which payment level the subscriber has chosen.
All users, regardless of tier, can access every Gemini AI model, including Flash-Lite, Flash, and Pro. The more advanced models are smarter but also consume more of a user's usage allowance. Each model additionally offers different thinking levels—Standard, Extended, and Deep Think—which affect response quality, speed, and how much of the usage quota is consumed.