My model quota is glitching. I got the pro subscription few days back. I did little bit of work, checked the usage and it showed 80% quota for the week remaining for 3.5 flash and 100% for all other models. The usage seemed very high for the little work I did so after few hours I checked again just to make sure but now the quota was down to 60%. I didn’t do any work in that period.
I am now returning after 3 days, my pc was shutdown the whole time and I see my quota is 0% for all models.
google doing a terrible work they should recruit me like i never seen a such bad work they did 5x worst their models got filter their models is less better than any models in any way and they their antigravity thing IS ABSOLUTE TERRIBLE : do they even test their product
Got the Ultra subscription thinking i was gonna get rid of this and now I waiting 2 days for damm thing to activate again. I refuse to buy more credit. This is ridiculous.
I moved from Gemini CLI to Antigravity, ran a few queries to look up the contents of an mcp server; usage was literally deleted within a handful of queries and now I have to wait 163 hours until it refreshes. How did I go from going all-day coding sessions to unusable? has there been ANY word on this problem from Google or are they just stoic about the whole thing hedging their bets people will go to Ultra to get 5 minutes more of uptime? This is incredulous.
Google is not giving what they promised when you pay for a plan.
The pro plan only allows for about four or five 5-hour block limits per week.
I would assume that a week would have 33.6 blocks of 5 hours.
Each day is supposed to have 4.8 blocks of 5 hours.
But you can hit the weekly limit after hiting the 5-hour limits 4 or 5 times in a week. And even worse, you still hit the 5-hour limit in less than 2 hours while using their lowest model.
I don’t understand what is the fuss they make about multiagents and parallel agents, if you can hit the limit using only 1 agent; no parallel tasks or anything.
If I were setting multiagent with 5 parallel tasks, I would only get 15 minutes of compute every 5 hours, but really only about an hour per week.
Today was my first time using Antigravity in about two months. I’m on the free plan and was only working on a small personal website project with some simple HTML, CSS, and PHP edits.
After just 1–2 coding tasks and roughly 5 minutes of use, my quota was completely exhausted. What confused me even more is that all available models immediately showed a refresh time of around 6 days instead of a short cooldown.
The requests were very basic website edits, not large codebase analysis, long conversations, or anything particularly demanding. A couple of months ago I could spend much longer working on similar coding tasks before hitting any limits, so this came as a surprise.
Has anyone else experienced this recently? Were there major changes to the free plan quotas, or does this sound like a bug with quota calculation?
I have noticed I do 3 times as much work with Claude code using sonnet 4.6 High and only hit the 5 hours limit and weekly quota goes down by just 10% while in antigravity there is no 5h hours quota, you exhaust your weekly limit after little bit of work with Flash.
Its main problem is Google, but also a problem is an inline context that every request has. I did a complete reverse engineering of Antigravity 2.0 and found very shocking things, and I posted the complete analysis on this forum, but the system deleted it. Another shocking thing that I found now is that when you are using Antiravity, you think you are chatting to Gemini 3.5 Flash, but internally it uses a smaller model like Gemini 2.5 Flash sometimes. After reviewing all things, I made a proxy by which you can use Anihravity 2.0, not just the IDE, with custom providers like Nvidia, OpenRouter. (GitHub - 12errh/antigravity-proxy: A TypeScript proxy that intercepts Antigravity's Gemini API and routes to NVIDIA or OpenRouter — chat, tools, browser automation, all working. · GitHub)
Its main problem is Google, but also a problem is an inline context that every request has. I did a complete reverse engineering of Antigravity 2.0 and found very shocking things, and I posted the complete analysis on this forum, but the system deleted it. Another shocking thing that I found now is that when you are using Antiravity, you think you are chatting to Gemini 3.5 Flash, but internally it uses a smaller model like Gemini 2.5 Flash sometimes. After reviewing all things, I made a proxy by which you can use Anihravity 2.0, not just the IDE, with custom providers like Nvidia, OpenRouter. (GitHub - 12errh/antigravity-proxy: A TypeScript proxy that intercepts Antigravity's Gemini API and routes to NVIDIA or OpenRouter — chat, tools, browser automation, all working. · GitHub)
I have the Ultra plan just renewed 2 days ago. Today I noticed using AI credits. This afternoon… locked up buy more credits. Seriously 280 dollars for two days?
Yep, it’s exactly like that. A couple of months ago I could code heavily for about 3 hours on the free plan before hitting the quota. Nowadays it’s gone after just a couple of prompts or minutes. From other posts it seems the Ultra plan is much worse now than the free plan used to be back in the day. Antigravity is a real bummer.
