So I have been a regular user of google pro on almost 2-3 Id’s , all under my name , and using antigravity for development and bug fixes. I might have been hit with 3-4 5 hours cool down this week , and this 5th cool down was directly of 7 days , directly from 5 hours to 7 days. it is just annoying and frustating at this point? how did I abuse the model , if I just did what was written in the documents.
I used opus 4.5 thinking because gemini 3 pro is trash for coding, instead of fixing the bugs , gemini 3 pro somehow removes the working part of the code , multiplying the bugs , so anyways i exhausted the limits of 4.5 thinking , I got a 5 hours cooldown. okay , now i start working next day. please tell me how I did I even abuse the limit? I was on a tour for last 3 days , where I did not even code , or touch antigravity at all , yet I got 7 hours cooldown limit?
please look into it , otherwise I might have to consider cancelling my google AI subscription to normal storage subscription , because I use no other features of the pro category
The recent surge in attention around Claude might have pushed Anthropic’s inference capacity to its limits, possibly leading them to throttle usage. It’s conceivable that their deal with Google means Google Antigravity users receive lower priority than paying Claude customers (just speculation here). Plus, neither the Google AI Pro nor Google One pricing pages mention any Antigravity quota in the subscription, let alone access to third-party models like Claude. Some clarity from Google would be helpful.
There is a bug in Antigravity that prevents you from logging in. Even though UI shows you as logged in, if you look in the Output tab → Antigravity, you will see it fails to log in. So basically you are “free user”.
I was waiting for a 42h coldown for anthropics models. Then on reset, it changed to 168h.
Great surprise from the 5h reset that was the main reason for buying PRO. WTF?
I might not need Opus or Sonnet most of the time. But how can they add a cooldown before even using it? They change the agreements retrospective is one thing, but this is another level.
I could have worked with 5d cooldown. Just need to install addons to keep track of it.
The main problem with Gemini models are not that they are inapt, it is that it takes so much time to teach them how to behave. They think they are the one who knows everything, and thus they become the guy in the team who listens to no one and assumes he is always right.
Often flash is better, but still suffer from the same bad system prompt.
I can work with it on some tasks. But it is really badly prompted for its system objectives. So, for critical correctness I would like to use the other models.
This is a real rug pull from google. But I hope they fix it. Lower quota, but still 5h reset, would be OK. At least then I know what I am working with.
This AI hype is just silly. Taking all our GPU’s, RAM and whatever to later not sell it to the agreed price.
I bought google ai plan just because antigravity was providing more limit for opus and a decent cooldown period of 5 hours before and now they are showing cool down of 5 days, i think its better to try claude code next month.
Im facing the same issues as well. I got a 7 day cool down limit on sonnet thinking. I have only used antigravity with claude for 2 days and they locked me out for 7 days. Its ridiculous
I have tried contacting google support for this and they mentioned that they cant do anything about it.
I thought it was an update mistake or error, but it seems to be the reality for all.
Gemini vibe code is good, but Claude is still better.
This cooldown period is not fun at all. Maybe there is a push for us to use Gemini only.
The Claude code UI/UX is not good. Specially when you have to do bugfix, it takes 3x more time than Antigravity… But if this goes on, i might be forced to use Claude code .
Hey, thanks for digging into this. I was worried I might need to upgrade to AI Ultra, but it seems like I can wait. I do have one question though. Even though I haven’t changed the plan or updated the version, I started getting a quota limit error two days ago, on Saturday morning PST, saying I need to wait around 3.5 days. Is there any fix we can apply on our end, or do we just need to wait for Google to resolve it?
Thank you for bringing this to our attention. We sincerely apologize for the inconvenience this has caused. We have escalated the issue to our internal teams for a thorough investigation. Any additional context or screenshots you can provide would be invaluable as we work toward a resolution.
My situation is even worse. Previously, my cool-downs were around 4-5 hours, just like @sbanthia75 mentioned. However, a few days ago, right after that time expired, it suddenly extended by another 4 days - until Jan 24, 2026, ~4:22 AM (!).
Today is Jan 24, 2026, and the cooldown timer just jumped to Jan 31, 2026, ~4:22 AM (!) without any action on my part. I haven’t used any Claude model even once during this whole time!
I also subscribed to the Pro account specifically to be able to use Opus or Sonnet occasionally, and right now I just feel cheated!
I just waited 7 days for my Pro plan quota to reset today (Jan 26th), only to be limited immediately after one single prompt. The Claude Sonnet 4.5 (Thinking) model consumed all available touches during its ‘planning’ phase. Now, I am forced to wait another 7 days just to use the service again. It makes the subscription feel useless.
I’m running into the same issue as of this week. Previously, the quotas for my Pro account had been working as advertised, refreshing every 5 hours. Starting this week, I noticed that my quota reset time was a few days out, even though I hadn’t even used antigravity for 4 days or so. This sucks, @Abhijit_Pramanik please make sure this issue is getting attention. Thank you sir!
