I am a Google AI Pro subscriber and I am confused about how quotas are being calculated in Antigravity.
I have been using almost exclusively Gemini 3.5 Flash Low. I specifically chose Flash Low because it is supposed to be a lower-cost model, and I have barely used the more expensive models.
Despite this, many of my other model quotas appear to be exhausted as well. Gemini 3.1 Pro (High) is currently unavailable and the dashboard shows:
Model quota exhausted
Refreshes in 4 days 12 hours
AI Credits: 0
My understanding was that Flash Low and higher-tier models would have separate limits, or at least that using a lower-cost model would not rapidly exhaust access to premium models that I barely used.
I would like clarification on the following:
Does Gemini 3.5 Flash Low consume the same shared quota pool as Gemini Pro High and other premium models?
If there is a shared quota pool, where is this explained to users?
Why is my quota refresh showing multiple days instead of the usual 5-hour refresh cycle?
Is this expected behavior, or could there be an issue with quota tracking on my account?
From a user perspective, it looks like using Flash Low is consuming the quota of models that I did not actually use, which is very confusing.
You’re obviously a new user which means you possibly took out the recent subscription deal? If so, they are secretly not honoring this deal for new users, you wont get the same quota as those paying full price…
This is exactly what happened to my account today, and it is a complete disaster.
I am also a paid Pro subscriber. I was mostly working with Gemini 3.5 Flash inside Antigravity IDE to write code, expecting it to be low-cost. But this broken unified shared quota pool system instantly drained my entire account.
Now, ALL models—including Gemini 3.1 Pro and Gemini 3.5 Flash—are completely locked out with a multi-day countdown timer.
Why am I being penalized and banned for days from using a… model just because the background agent kept looping workspace files? Google needs to decouple the Flash quota from the Pro High tier immediately, or stop selling this as a “Pro” subscription. Fix this entitlement sync and reset the stuck multi-day counters for paying users!
As a paying subscriber, I cannot recommend Gemini in its current state.
Google offers multiple paid tiers, including AI Pro and AI Ultra, creating the expectation that paying customers will receive a reliable and transparent experience. However, the shared quota system completely undermines that expectation.
If using Flash, Pro, and other Gemini models all drains the same quota pool, then users are effectively punished regardless of which model they choose. There is little incentive to use lighter models if they still contribute to locking access across the entire Gemini ecosystem.
What makes this worse is the lack of transparency. Most users discover the true limitations only after their quota is exhausted and they are faced with multi-day lockouts despite being paying customers.
A paid AI subscription should not leave users unable to access core features for days after normal usage. Google needs to provide clear quota reporting, separate quota pools for different model classes, or significantly higher limits for paying subscribers.
Until these issues are addressed, I cannot recommend Gemini to developers or professional users. For many of us, it has become a source of frustration rather than a productivity tool.
Paying customers deserve clarity, predictability, and access to the service they are paying for.
