For the first time ever, I hit a weekly quota on a paid plan? And the shocking part is, I didnt even push it that hard? 5x 5h quota windows and suddenly im shooting blanks, after just 3 days? Idk if it’s a bug, but how does this happen on an Ultra plan? I never even had this issue on Pro, let alone on Ultra and I barely even touched 3.1 pro, 99% of the work was done via 3.5 flash?
Everyone is having the same issue, probably a disgruntled dev or industrial sabotage Friday afternoon before leaving for the weekend, otherwise Google just blew up its AI business model, who knows, will leave and cancel my sub if it happens again next week
Yeah same, after Friday finishes, weird stuff.
It’s definitely not normal, I mean ultra $100 is supposed to have 5x Pro quota, it shouldnt even be possible to hit weekly quota in 3 days? Idk, they seriously need to fix this and fast, because I switched to ultra to get work done, not get locked out for half a week. If this is what it’s gunna be, I’m going to have to switch to cursor or codex. They cant expect people to be satisfied with this. And they really need to fix their dev pipeline. If I pushed … updates that werent properly tested, like this… I’d probably be out of a job by now. I just dont understand why they cant open a canary channel like everyone else, where they give you a slight increase in quota and latest updates, but 100% consent to providing logs and feedback, so the people just doing work can actually help them test it, instead of finding out once 100k+ developers are locked out… This should have been caught before it got released and this should have been dealt with by a beta channel.
Same issue for me, even if you disconnect / connect again. 2.0.4 version. Not a single email to warn about a ‘small’ change like this ?
Reason here :
No changelog either. Which is the real problem. We need clarity of what, why, when they make changes to the systems we use.
Yes .. they have nerf’ed it again.
The thing for everyone to remember is this: if they were truly interested in transparency, they would give you tokens and tell you how they are used. Then when you ran low, you could buy more tokens. They are intentionally obfuscating the process to make it easier to deceive us. They don’t want you to know how it works.
Clear == a token cost per submission with details if you want it.
Sneaky == the inability to know how much you really have left and where it all went.
There is no reason not to do the first one .. unless you don’t want people to know how you are using their money.
Same problem for me. I have never run out of quota for serval days.
I don’t know what happens. It seems the refresh period is 2 days now. Gemini Pro user.
Especially because you can see Vertex AI token usage, when you wire in a GCP. You can check your billing, but antigravity doesnt say.
Cant be, cuz mine got locked for 3 days 9h
I switched to 1 of my free accounts, to check. Literally 1 prompt on claude, it doesnt even finish (fresh chat) and 2 prompts on gemini 3.5 flash (low) and account is locked for a week…
Same here, my quota ran out today with very little use. They don’t know what they’re doing to monetize the models. Crazy world.
It is showing that my Claude models have 80% usage, while I have not used Claude models in the last 15 days. Where were the credits use, since I now exclusively use Gemini 3.1 Flash (High). Still it is showing that credits are used. So my question is that where are the credits used. Am on AI Pro plan.
same here, im just using a single gemini model but the other gemini models are getting exhausted too. its pretty useless at this point