Severe Token Drain After Just a Few Prompts

What the hell is the actual weekly limit? And how many tokens are we even getting in that 5-hour limit when sending just a few prompts is enough to trigger the 5-hour lockdown?! How on earth does hitting the hourly limit just 4 times slash my remaining weekly limit straight down to 50%?? That means hitting the hourly limit a mere 8 times completely wipes out the entire week’s quota! If that’s the case, calling this a ‘weekly limit’ is a blatant lie when it can’t even survive 3 days of real-world use!

What is the damn point of paying for the Google AI Pro package then?? Are you guys actually running a transparent business, or are you just treating Antigravity users like a total joke? Do we look like clowns for you to play with and scam?!

To top it all off, today I got hit by a massive bug where every single model got stuck on ‘Working…’ and BOOM—just like that, I was slapped with a 5-hour lockdown and my tokens were completely drained to zero! Do you people even bother to test your product before shoving an update out? Because if you don’t, this is shockingly unprofessional. This is worse than the work of an amateur apprentice developer, let alone an entire team working for a supposedly top-tier, global tech giant!

Take a close look at the text in my screenshot. Your dashboard explicitly claims that ‘limits will last longer with shorter tasks or using more cost-effective models’ and that the 5-hour limit exists to ‘fairly distribute global capacity across all users.’ What a hypocritical statement!

How is it fair when a severe system bug gets stuck on ‘Working…’, loops endlessly, and strips away a paying user’s entire 5-hour quota to 0% in the blink of an eye? You state that the ‘weekly limit is tied directly to your individual tier.’ I paid for the Google AI Pro tier, yet my Weekly Limit is already hacked down to 49% with 6 days and 11 hours left to wait.

Your system failed to deliver the promised ‘longer-lasting limits’ for my tasks. Instead, your broken background processes consumed my quota proportionally to your own technical failure, not my actual usage! This completely invalidates the ‘fair distribution’ policy you boast about on this very screen."
Let the math sink in for a second: I have only used this system for a single day, and 51% of my entire weekly limit is already completely wiped out! Just ONE day, and more than half of my weekly quota has vanished into thin air because of your broken system!

Look again—your dashboard literally tells me that I have to wait another 6 days and 11 hours just to get a refresh. Are you out of your minds? If 1 day of minimal use eats up 51%, it means your so-called ‘weekly’ limit cannot even survive 2 days under normal circumstances!

This isn’t just a minor miscalculation; this is a catastrophic failure of your quota system. You are heavily throttling paying users right out of the gate. How do you expect anyone to rely on your platform for actual work when your ‘weekly’ service burns out before the second day even begins?!

Can you share some info about what you are prompting for?

I hit limits every now and then. What you describe seems extreme though. What kind of tasks are you typically running? Are these single prompts or multi-turn sessions? And when you say tokens got drained to zero, was that tied to the bug you mentioned, or separate instances? Might help clarify what’s happening.

just stay at “Working…” and nothing happend then boommmmm. They said just like this !

(300) Angtigravity BUG !!! - YouTube