Enterprise technology teams integrating Google’s Gemini AI into workflows now face new usage limits based on computing power rather than the number of requests. According to WIRED, Google rolled out these changes earlier this summer as part of broader AI upgrades, and the new metering system applies across all Gemini tiers: Free, Plus, Pro, and Ultra. Organizations relying on Gemini for automation, document summarization, or custom coding must understand these limits to avoid service interruptions.
How AI Usage Is Changing
Google now measures Gemini AI usage by the computing power requirements of your requests, not by the number of requests you make. WIRED explains that this means generating two complex videos might deplete your quota faster than three simple ones. This shift makes it easier for Google to manage data center costs but leaves users less certain about when they will hit their limits. Additionally, Google states in its support documents that “access is subject to change or may be limited based on testing, experimentation or availability.” WIRED notes this effectively means some days may allow more usage than others.
The two main factors determining usage are the plan you are on and the complexity and length of your prompts. The chosen Gemini AI model—for example, Flash-Lite, Flash, or Pro—also affects consumption. Each model has different “thinking” levels: Standard, Extended, and Deep Think, which impact response quality, speed, and usage limits.
The Quotas for Each Plan
Google offers four subscription plans for US users. The table below summarizes pricing and relative usage allowances:
| Plan | Monthly Price | Usage Limit Relative to Free | Context Window (approximate words) |
|---|---|---|---|
| Free | $0 | Standard (unspecified) | ~24,000 words (32K tokens) |
| AI Plus | $8 | 2x standard | ~96,000 words (128K tokens) |
| AI Pro | $20 | 4x standard | ~750,000 words (1M tokens) |
| AI Ultra | $100 or $200 | 5x or 20x of AI Pro | ~750,000 words (1M tokens) |
WIRED reports that Google does not specify exact limits on the free tier, describing them only as “standard.” Higher tiers provide multiples: 2x on Plus, 4x on Pro, and either 5x or 20x on Ultra depending on payment level. All users can access all models, but more advanced models count more toward usage.
Check Your AI Usage
To monitor your current usage, WIRED outlines the following steps:
- Web app: Click the cog icon (lower left) then Usage limits.
- Mobile app (Android/iOS): Tap the menu button (top left), then the cog, then Usage limits.
You will see two bars. The top bar shows your current usage, which resets every five hours. If you exhaust it, you must wait until the next reset; the app indicates the reset time.
For enterprise decision-makers, these changes mean that budgeting for AI services requires estimating compute demand per prompt rather than request volume. The vague “standard” limits on the free tier and the variability from experimentation make it essential to test workloads against each plan. Organizations using Gemini for high-volume tasks like data analysis or automated customer support should consider the Pro or Ultra tiers to ensure adequate capacity—and should plan for potential fluctuations in available quota due to Google’s experimentation policies.
Sources:
