Is anyone else completely fed up with the ridiculous quota limits and reset timers on the Antigravity platform? I am hitting a massive wall here, and it’s completely killing my development speed.
Take a look at the screenshot I’m attaching. My Gemini 3.1 Pro (High) quota is sitting at 20%, and the reset timer is telling me I have to wait 35 hours! How is any developer supposed to get serious work done when the primary model is effectively locked down for a day and a half?
But here is the most infuriating part: I haven’t even used the platform for the past few days! How on earth is my quota sitting at 20% when it has been completely idle? Is there a bug in how usage is being calculated, or are quotas silently draining in the background?
I am paying for a Google AI Pro subscription specifically to avoid these exact bottlenecks. I am currently in the middle of a massive Next.js redesign for a major project, which requires heavy, continuous iteration and debugging. I cannot just put my entire production schedule on pause until tomorrow night just because the quota algorithm is acting up.
Also, it’s frustrating that Claude Sonnet/Opus 4.6 and GPT-OSS 120B are sitting right there at 60% availability. Why is Google’s own flagship model the most heavily restricted (and seemingly bugged) on their own agentic platform?
We need a transparent explanation of how these Antigravity quotas are calculated and an immediate fix for these ghost-drains on our limits. A 35-hour reset timer for an unused Pro account is an absolute joke.
Has anyone else experienced their quota draining without even using it, or found a way to escalate this?
I agree. I do not even mind that we get rate-limited, and as this is causing real world costs, we should be rate limited, I think. (Yeah, I know this sounds funny. But this is the only way we keep striving to achgieve the ultimate goal, which would be affordable and yet “performant-enough” local AI)
However, it is totally unacceptable that the quota usage is so intransparent, and even worse, the totally random “refresh” policy. Sometimes you suddnly get hit with what feels like a “ban”, and what software developer would be willing to accept this? It’s like a car that sometimes works, sometiomes it doesnt. Thats ok for a hobbyist and his vintage car, but not for a cab driver.
I am not even considering upgrading from “Pro” to “Ultra”, because who knows when , if and for how long I’ll get rate limited there? This intransparency is seriously hindering antigravity to become a platform for real work.
all those expensive subscription just so I cant even finish a simple project development ? Yeah, why dont u just put ads bar onto our window ? And charge us more on every prompt we gave. If u really need that money, I’ll spare u some change so that u can keep those AI clanker servers running.
I have a PRO AI plan, and since this morning the Gemini Pro model has 24/7 recovery… but I used it yesterday without any problems. Have they changed anything or are there any problems? I’m working on it! If I don’t get it fixed, I’ll go back to Claude Code.
I thought this was a bug. Yesterday my quota dropped from 60% to 0% without any prompt, and the worst part was seeing the reset time set to 50+ hours. I assumed it had to be a bug and hoped I’d be able to work today.
I basically didn’t use the tool at all this week. I really hope this isn’t working as intended.
Just came out of a 5d lockdown, used the tool for 2 days, hitting 5hr lockdown once→ just lost the recent conversations after upgrading→got hit with new 5d lockdown.
This is insane for a paid service.
Title: [Pro Sub - Quota Bug] Gemini 3.1 Pro Quota Suddenly Dropped to 0% with >1 Day Reset Time
Hello team,
I’ve actually had the same issue with quotas. Recently, I have encountered a severe anomaly with the quota tracking in my environment. My Gemini 3.1 Pro quota unexpectedly plummeted from around 60% directly to 0% without any corresponding heavy usage on my end.
Observations & Steps Taken:
The Quota Monitor currently shows 0.00% for Gemini 3.1 Pro, with an abnormally long reset timer of 1d 4h 52m.
Interestingly, my Gemini 3 Flash quota remains completely unaffected at 100%.
I have already tried completely clearing the local .antigravity cache folder via PowerShell, but the locked state persists.
please investigate this account synchronization issue and assist in restoring the proper Pro quota state asap.
I invalidated the auth token in google security, logged out and back in and now I have full quota’s again randomly, but my refresh window went from 133 hours to 167?
Unfortunately, my and many other AG users’ model screens aren’t live. I have to close and restart AG to update my model quota screen. Talk about a workflow-crushing feature!
My workflow is: prompt, review, restart AG to check model quota, then prompt again. Forget multi-agent workflows.
Yet, using Gemini CLI in an AG terminal works fine and doesn’t hit quota limits.
This morning after logging out and back in again, the quota shows full, but it hits the limit immediately. Also the refresh time is unchanged (±4 minutes) Their quota tracking is a complete mess…
