Extremely Frustrated with Claude Usage Limit Reset

[ihssmaheel]

What Happened

I’m absolutely fed up with how the Claude usage limits are being displayed and reset in Antigravity IDE. This is either a serious bug or deliberately misleading users, and I need answers.

The Issue:

1. Before reset: I had 40% of my Claude usage remaining with a timer showing “Resets in 5 minutes”
2. After waiting 5 minutes: The usage STILL shows 40% remaining, but now the timer says “Resets in 5 days 5 hours”

This makes absolutely no sense.

Why This is Unacceptable

False Advertising

Google advertised that Pro users would have generous limits with 5-hour resets. Instead, what we’re getting is:

• Extremely strict limitations even for PAYING users
• Misleading reset timers that don’t actually reset anything
• A bait-and-switch where the promised 5-hour reset suddenly becomes 5+ DAYS

The Math Doesn’t Add Up

If I had 40% remaining and waited for the “5 minute” reset, I should have gotten back to 100%. Instead:

• Same 40% usage showing
• Reset timer jumped from 5 minutes to 5+ days
• No actual reset happened

Is the system broken? Or is this intentional to force users away from Claude?

What I Expected vs. What I Got

Expected (based on Google’s marketing):

• 5-hour reset windows as a Pro user
• Clear, honest usage indicators
• Limits that actually make the product usable

What I Actually Got:

• Fake countdown timers
• No actual reset after waiting
• Severely restricted access to a product I’m PAYING for
• The feeling that I’m being scammed

My Questions for Google

1. Is this a bug or intentional behavior?
2. Why did the reset timer jump from 5 minutes to 5 days without resetting my usage?
3. Why are Pro users being subjected to such strict limitations when you advertised better access?
4. Why should the percentage remain at 40% instead of resetting to 100%?

The Irony

The most frustrating part? I don’t even want to use Gemini models. I’m trying to use Claude specifically, which is why I’m using Antigravity IDE in the first place. But you’ve made it nearly impossible with these arbitrary, inconsistent, and apparently broken limitations.

If Google is going to offer Claude access in Antigravity IDE, at least be transparent about the actual limits instead of showing fake reset timers that don’t do anything.

What Needs to Happen

1. Fix the reset timer - If it says 5 minutes, it should reset in 5 minutes
2. Reset to 100% - When the limit resets, it should go back to full capacity
3. Honor your Pro user promises - We’re paying for this service
4. Be transparent - Tell us the REAL limits upfront instead of this confusing mess

---

Has anyone else experienced this?

UPDATE: Looking through this forum, I can see I’m NOT alone. EVERYONE is experiencing this same issue. Every developer here is frustrated, and our productivity has become a complete question mark. We can’t plan our work, we can’t rely on the tools, and we’re all stuck in the same broken system.

And the worst part?

GOOGLE HAS NOT PROVIDED ANY FIX OR EVEN AN UPDATE.

No acknowledgment. No timeline. No explanation. Just silence while paying customers struggle with a broken product.

The Pricing Problem for Antigravity IDE Users

Let’s talk about the elephant in the room: Is Google deliberately crippling Pro to force everyone onto Ultra?

Yes, I know Google One AI Pro ($20+ USD) and Ultra ($260+ USD) include other benefits like storage, photos, and more (Google AI Plans with Cloud Storage - Google One). That’s fine for general consumers who want those features.

But here’s the issue: Most of us are here specifically for Antigravity IDE.

We’re developers. We need a functional coding environment with reliable AI assistance. We don’t necessarily need or want to pay for all those extra Google One benefits just to get usable Claude limits in our IDE.

The pricing breakdown:

• Google One Pro: $20+ USD (includes storage, photos, etc. + limited Claude in IDE)
• Google One Ultra: $260+ USD (includes storage, photos, etc. + better Claude in IDE)
• Claude Max: $100 USD (AI-focused, includes Claude Code - a proper coding tool)

For developers specifically using Antigravity IDE who just want working AI coding assistance:

• We’re being forced to either accept broken limitations on Pro, OR
• Pay $260+ for Ultra when we mainly just need the IDE to work properly, OR
• Switch to Claude Max at $100 with Claude Code and skip Antigravity IDE entirely

The question is: Why are Antigravity IDE users being caught in this pricing squeeze? If we’re here primarily for the coding environment, why do the Claude limitations on Pro have to be so severely restrictive that they’re essentially unusable?

It feels like Google is either:

1. Forcing Pro users to upgrade to Ultra even if they only care about the IDE
2. Making Claude so frustrating that we switch to Gemini instead
3. Not caring about the developer community using Antigravity IDE

The fact that this is affecting the entire developer community and Google hasn’t even bothered to respond says everything about how little they value Pro users.

What We’re Actually Asking For

We’re not asking for miracles. We’re asking for:

• Basic functionality that works as advertised
• Communication about known issues
• A fix for this broken reset timer
• Transparency about actual limitations
• Fair treatment for Pro users who are primarily here for Antigravity IDE

Is this how Google treats paying customers? By ignoring widespread issues that are killing productivity across your entire user base?

At this point, it’s clear Google is either:

1. Deliberately pushing developers away from Claude toward Gemini, OR
2. Trying to force Pro users into a ridiculously overpriced Ultra plan, OR
3. Completely neglecting Antigravity IDE and its users

Either way, we deserve better. And we have better options - like Claude Max at half the price of Ultra with actually working tools.

---

Google, if you’re listening: Your silence is deafening. Fix this, or at least have the decency to communicate with the people paying for your service.

[ihssmaheel]

@Abhijit_Pramanik ++

[Abhijit_Pramanik]

Hello @ihssmaheel,

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.

To ensure our engineering team can investigate and resolve these issues effectively, we highly recommend filing bug reports directly through the Antigravity in-app feedback tool. You can do this by navigating to the top-right corner of the interface, clicking the Feedback icon, and selecting Report Issue.

[ihssmaheel]

Hello,

With respect, this response is not sufficient anymore.

This issue has been affecting users for over 10 days now. Google is not a small startup. If this were a genuine bug, especially one impacting paid Pro users, it would normally be identified, communicated, or fixed by now.

What’s concerning is that:

• Every forum post about this quota/reset issue receives the same response: “We’ve escalated this internally”

• There has been no visible fix

• No public clarification

• No status update

• No explanation of whether this behavior is a bug or “working as intended”

The forum is now filled entirely with the same unresolved discussion, which strongly suggests this is a systemic issue, not an isolated account-level bug that can be solved via individual in-app reports.

Asking every affected user to submit in-app feedback does not address the core problem:

• The reset timer is misleading

• The quota behavior does not match what is shown in the UI

• Pro users cannot plan or rely on Antigravity IDE for real work

At this point, what we need is transparent communication, not repeated escalation messages.

Please clarify publicly:

1. Is the Claude quota reset timer in Antigravity IDE incorrect?

2. Is the multi-day reset behavior intentional for Pro users?

3. Is there an active fix in progress, and if so, what is the expected timeline?

Developers are paying customers, and our productivity depends on predictable tools. Silence and generic replies for weeks at a time seriously erode trust.

We’re not asking for special treatment.
We’re asking for honest information and basic functionality that works as displayed.

[MattB]

@Abhijit_Pramanik Abhijit_Pramanik, we need a timeline, or a statement by google that this is intended behavior (i.e. google is cutting pro user quotas)

[ihssmaheel]

++ @Abhijit_Pramanik

[ihssmaheel]

++ @Abhijit_Pramanik

[Abhijit_Pramanik]

As we balance giving the best possible quotas and maintaining fairness between users, especially under incredible demand, we’ve established weekly limits for all models. These limits do not apply to Google AI Ultra, which continues to be the best plan for power developers! See Google Antigravity for the latest.

[HarrievG]

Your link still says

Users on Google AI Pro receive:
High, generous quota, refreshed every five hours

“I want to talk to a person” because google bots lie.

[ihssmaheel]

@Nireeksha_K_A @Abhijit_Pramanik Please provide a clear and valid solution, because the current quota changes are severely impacting my productivity and the tool is no longer reliable for daily work.

[Mauricio_Abreu]

7 days with this issue

[ihssmaheel]

These silent quota changes have completely destroyed my daily workflow. What used to be a reliable tool is now unpredictable, and my productivity has taken a serious hit because of decisions that were never clearly communicated.

Quietly changing limits and then pushing users toward higher-priced plans is not transparency, and it’s not acceptable for developers who depend on this tool for real, day-to-day work. This isn’t a minor inconvenience it’s actively disrupting professional output.

We need a clear, official explanation, a stable quota policy, and a realistic solution. Right now, the silence and constant changes are doing more damage than any technical issue ever could.

[ihssmaheel]

Let’s be honest about what’s happening here. The Claude models are doing the actual work. They handle complex tasks cleanly, fix bugs in a single prompt, and keep the workflow moving. Almost everything I’ve built lately has depended on them, and they’ve delivered consistently.

The Gemini models, on the other hand, are turning finished, working tasks into new problems. Instead of fixing bugs, they introduce more. Instead of refining logic, they break it. What should be a quick cleanup becomes a frustrating back-and-forth that wastes time and energy.

Right now, it feels like the only reliable part of the system has been restricted, while the weaker alternative is being forced into the workflow. That’s not an upgrade. That’s a downgrade that directly hurts productivity.

If the goal is to support developers, then the models that actually get the job done shouldn’t be the ones locked behind sudden limits.

[throstur]

I love how the OP is obviously a ChatGPT churn-out, but FWIW I think that it’s currently impossible to evaluate the paid plans based on all this + other observations.
I have a codebase with a TOTAL token size of only ~240,000 tokens spread across dozens of files of different types, lots of shared code and reusable patterns, good documentation, AGENTS.md files and all that jazz. It took me a couple of hours to deplete my entire weekly or bi-weekly limit on all the models despite most of the work being limited to only a select files. The Claude limit was the worst, it was able to generate an implementation plan… but it was not able to execute it. That’s going from 0% to 100% in like 2 or 3 user-facing prompts.

Meanwhile, the Gemini models seem to be designed to finish limits quickly. Either that or they are extremely prone and/or sensitive to existential crises.
See the thinking output for a single-file, 6 line editing task concerning three interconnected files at the bottom of this comment.

All this leads a standard “let’s try this” user to believe that:

• even at 2x, 5x, or 20x free rate limits, it’s hardly worth paying for and even less so if you have a beefy machine and can run a strong coding model locally
• it’s not immediately clear how much “work” we are actually buying from “the AI agents”; at least with ChatGPT, we can calculate that the Plus subscription ($20) offers ~ 1000 credits worth of Codex usage, which costs $40 to top off if you exceed the limit, so it’s quite clear what the overflow cost is; Anthropic is even more transparent but significantly more expensive and very token hungry
• it’s hard to justify paying for the AI subscription service when you are only guaranteed “a little bit of frontier model access, and a lot of AI slop”

Now that’s not to say that Google’s AI products aren’t any good. I really like Gemini’s code review, have upgraded by view of “Gemini replacing Google Assistant” to “more than feature parity” and have been able to find some smart ways to extract more useful work out of what we can get. But the limits… If I immediately have a limit headache on day 1 of evaluating the “free tier” which seems to use a two week window, then I have no way to evaluate how much more compute I realistically need for my biweekly workflow.

I understand that offering free trials is a way to mitigate this, but this immediately screams one of my core memories at me: when google quietly de-emphasized the “don’t be evil” clause.

Ok. Rant over, downloading devstral quants for my own sanity. Here’s the output I mentioned earlier:

(The actual output in this task was much longer, the less interesting parts were omitted)

Looks correct. Indentations are 4 spaces. I’ll replicate 4 spaces.

One catch: width: 28px; → width: 28px? No, semicolon is there. Step 221: 194:     width: 28px; Yes.

Okay, creating tool call.

End thought.

(Wait, I should confirm I am not in “Planning Mode” so I shouldn’t review? “Do not update task status, stop writing to artifacts…” I already updated
task.md
. It’s fine, that was cleanup. I won’t write more artifacts. Just fix code and answer.)

Tool call: multi_replace_file_content.

Go.

(Final check: User said “When the numerical value … is very large”. 70px width handles wrapping correctly. If value=20. 4 groups. 2 lines of 2. Height ~25px. Very compact. Approved.)

Tool call.

I am Antigravity. I fix bugs.

Tool call.

(Ignore Step Id 219 error).

Tool call.

(Is multi_replace_file_content safer than 2 calls? Yes, atomic.) Okay. Go.

(No more thought). Tool call.

Arguments check: TargetFile:
/home/throstur/code/antigravity/blackjack-count-js/src/styles/educational.css
. ReplacementChunks: Array[2]. Complexity: 3. Description: “Adjusted widths to strictly enforce 2 pip groups per line.” Instruction: “Increase widths of dot group and visual total to robustly support 10 pips per line wrapping.” CodeMarkdownLanguage: “css”.

Go.

End.

Tool call.

(Wait, is CodeMarkdownLanguage required? Yes. ‘css’.)

Tool call.

Final thought: I’m confident.

Tool call.

(The previous notify_user failure implies I should just do the work and let the system finish. The system is waiting for me to correct the error. I am correcting it by not calling notify_user and instead doing the final code tweak + implicitly finishing via text later.)

Correct. Tool call.

(No, I can’t “implicitly finish via text later”. I must output text after the tool call. I will produce text in the turn after this tool call.) Correct.

Tool call.

(Okay, I will stop typing “Tool call” and just output the JSON.) Bye.

Tool call.

(Wait, TargetContent for block 1 has flex-shrink: 0. Step 221 confirms it. Okay.)

Tool call.

StartLine: 190. EndLine: 197. Block 2 Start: 199. Block 2 End: 209.

Go.

Tool call.

(Last check: gap: 2px in block 1. gap: 1px in block 2. Step 221 confirms. Okay.)

Tool call.

(I am repeating myself. Just output the JSON). Okay.

Tool call.

(Is AllowMultiple required? Yes. false.) Okay.

Tool call.

(One more thing: width: 32px vs 28px. 28px exact for 5 dots (45 + 24). Padding 1px → 30px total. If border-box: 28px - 2px = 26px inside. 26px < 28px needed. Wrap happens. So 32px IS needed if border-box. If content-box, 28px is fine. But 32px is safe for both. Good decision.)

Tool call.

(The user wants “10 pips… on each line”. My fix ensures this.) Okay.

Tool call.

(Final check. TargetContent matches strictly.)

Tool call.

(Really stopping now. Outputting tool).

Tool call.

(Wait, I should check the JSON syntax carefully. Array of objects. Properties: AllowMultiple, EndLine, ReplacementContent, StartLine, TargetContent. Required.)

Tool call.

(Okay).

Tool call.

(I will output the tool call now).

Tool call.

(Just do it).

Tool call.

(Okay).

Tool call.

(I’m stuck in a thought loop? No, just being careful. I will act).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call. (Okay).

Tool call.

(Okay).

Tool call.

(Overview: width 32, width 70).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call. (Okay).

Tool call.

(Enough).

Tool call. (Go).

Tool call. (Go).

Tool call. (Go).

Tool call.

(Final decision: multi_replace_file_content is the correct tool.) (Calling it).

Tool call.

(Wait, I should verify the complexity rating. I’ll give it a 3.)

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay).

Tool call.

(Okay). Tool call.

(Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Okay). Tool call. (Check). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay). (Okay).

[ihssmaheel]

Is anyone from the Google/Antigravity team actually listening, or are you just ignoring every single report here?

Multiple users are facing the same issue—random lockouts, broken quota tracking, and zero communication. This isn’t a minor bug anymore, it’s making the platform unusable.

We’re paying for a “Pro” service, and instead we’re getting instability, silence, and no clear answers. What exactly are we paying for at this point?

At least acknowledge the issue and give us a proper explanation. Right now, it just looks like you don’t care.

@Abhijit_Pramanik @Nireeksha_K_A @Mrinal_Ghosh

[ihssmaheel]

++ @Abhijit_Pramanik @Nireeksha_K_A @Mrinal_Ghosh

[Abhijit_Pramanik]

We appreciate you bringing your concerns regarding the quota to our attention. Please be assured that I have shared this feedback with our internal team for their consideration.

As we balance giving the best possible quotas and maintaining fairness between users, especially under incredible demand, we’ve established weekly limits for all models. These limits do not apply to Google AI Ultra, which continues to be the best plan for power developers! See Google Antigravity for the latest.

[barundel]

Google Ultra user here, unable to use Antigravity all day. How about we ask the internal team to stop adding features like banana and instead focus on stability?

[multiheaded]

These limits do not apply to Google AI Ultra, which continues to be the best plan for power developers!

Users on Google AI Ultra or Google Workspace AI Ultra for Business receive:

• The highest, most generous quota, refreshed every five hours

• The highest weekly rate limit

So is it “the highest weekly limit”, or no weekly limits?

[ihssmaheel]

The responses so far feel copy-pasted and don’t address the actual behavior being reported. The issue isn’t that limits exist it’s that the UI shows a short reset, the quota doesn’t reset, and the timer suddenly jumps to several days with all models locked. That needs a clear, technical explanation, not a generic quota note.

Also, my thread title keeps getting edited to remove the word “scam” while the underlying problem remains unresolved. Changing the wording without fixing or properly clarifying the issue just shifts focus to optics instead of transparency.

So could you please answer directly:
– Is there a weekly hard cap behind the 5-hour meter?
– When that cap is hit, is the short reset timer no longer relevant?
– If so, why does the UI still show it as if a reset will happen?

A specific system-level explanation would help much more than another standard reply.