[Feedback] Critical Concerns Regarding Ultra Quota Reductions, Loss of Transparency, and Broken Promises

To the Antigravity Team and fellow builders,

As a paying subscriber to the Ultra plan, I am writing to express deep concern regarding the recent shifts in quota management. We upgraded to Ultra because we require stable, predictable, and high-capacity resources for our production environments. However, recent unannounced changes and broken promises are severely undermining our trust in the platform.

The “Boiling Frog” Experience: A Timeline of Downgrades

What troubles developers the most isn’t just the reduction in available resources, but the deliberate removal of data transparency during this process. The recent changes feel like a calculated downgrade:

  • The Loss of Metric Transparency: Previously, we could see our exact Token and Request counts. This was suddenly replaced by an obscure percentage (%). The Impact: It is now impossible for developers to monitor exact usage, calculate costs, or properly manage concurrency.

  • Massive, Silent Quota Cuts: After masking the exact numbers behind a percentage, the actual underlying limits were noticeably reduced. The Impact: Applications are frequently hitting rate limits, which directly breaks production stability for agentic workflows.

  • Zero Communication: These sweeping changes to production-critical resources were rolled out silently, without any prior announcement or changelog.

The Blatant Contradiction: Official Terms vs. Reality

To make matters worse, there is a glaring contradiction between what is officially promised to Ultra users and what is currently being delivered. According to the official documentation for the Ultra tier:

(Please see the attached screenshot)

The terms explicitly state we should receive:

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

  • No weekly rate limit

The Reality: We are dealing with an opaque UI, massive hidden cuts, and many users in this forum are hitting invisible walls that certainly do NOT refresh every 5 hours (some are waiting days, as seen in other threads).

If the official documentation still states these terms, but the actual service has been silently throttled to a fraction of this promise, this goes beyond a lack of transparency—it borders on false advertising.

Our Constructive Demands

As core users who financially support Antigravity, we want the platform to succeed, but we need it to remain developer-friendly and honest. We urgently request the official team to address the following:

  1. Restore Metric Transparency: We strongly request the immediate return of exact Token/Request counts in the dashboard. Vague percentages are useless for serious development.

  2. Honor Your Documentation: Address the severe discrepancy between the documented “5-hour refresh / no weekly limit” and the actual throttled experience developers are facing right now.

  3. Publish Clear Quota Rules: Stop the silent updates. If capacity constraints require adjustments, publish an official changelog detailing the exact numerical differences.

Until this lack of transparency is addressed and an official response is provided to clarify these broken promises, it is incredibly difficult for us to confidently build on or commit further financial resources to this platform.

We look forward to an official clarification from the team.

24 Likes

This seems to me like a global reduction in AI usage and not just on ultra.

Me too, I spend money to buy a service package worth the money spent, however the quality of the model is decreasing day by day, and the quotes are increasingly being restricted. I hope you have a way to solve this problem.

3 Likes

To be completely honest, if this is how it’s going to be, I’m done with this Ultra product. I’d rather just go to Codex.

6 Likes

Was about to say the same. I have tried Codex this week and overall the AI is a better at interpreting prompts and you can see what the quotas left are.

YEH! THIS IS NOT RIGHT! Whats going on!

I’ve been putting up with this ever since the Professional plan. Every gradual quota reduction, like boiling a frog in warm water, has been met with zero official response. Their attitude is plain and simple: complain all you want. We’ll keep cutting the quota and never restore it. After all, they’ve already rolled out AI credits.

4 Likes

Hello @LOVECHEN @aaron_wu @mingcheng_mingcheng @Jonathan1 @Barry_Bolton @bambooz0412!

Thank you for bringing these concerns to our attention. Please be assured that I have shared your feedback with our internal team for further review.
We appreciate your continued patience as we work to enhance the Antigravity experience.

I really hope your colleagues do something. A lot of us, the users have complained in the last 2 months and nothing happened. It just got worse

3 Likes

Even if they found a bug or something, the lack of transparency completely broke my trust. I’m very disappointed. ‘Vibe coding’ wasn’t something I believed in at first, but AG accelerated my projects by months or even years. I’m grateful for what I was able to achieve, but it’s sad to see it end like this :confused:

3 Likes

Hi @Abhijit_Pramanik,

Thank you for stepping in and passing our feedback to the internal team. We really appreciate the communication.

I want to emphasize that the intense frustration in this thread comes from how much we rely on Antigravity for our actual production environments. The sudden shift to the opaque 5-tier percentage system and the unexpected quota walls have completely removed our ability to predict costs and manage our workflows reliably.

As you can see from other developers mentioning alternatives like Codex, this lack of transparency is making it very difficult for teams to justify staying on the platform right now.

While the team reviews this, could we please get some clarity on two critical points?

  1. Transparency: Is there any roadmap or plan to bring back the exact raw token counts (input/output) to the dashboard so we can optimize our agents responsibly?
  2. Quota Clarification: Can the team transparently confirm whether the actual compute limits for the Ultra plan have been adjusted compared to the original documentation?

A rough ETA or any technical insight on these two points would go a long way in rebuilding trust with the developer community here. Thanks again for your help.

Hi Abhijit, Appreciate the response, but what exactly are you expecting to come of this escalation, if anything? This is a complete sham and stopping people doing their work and seem completely not thought out.

2 Likes

No, it really isn’t. They just keep shifting it. Now it’s in Google AI Studio.

1 Like

whats funny to me is how people are using ai to complain and explain how their ai doesn’t work.

”Boiling frog” might make sense to you and your AI, but here it looks like your “vibe isnt good”.

2 Likes

Yes, I have similar issues and concerns on the Pro package. It has basically become unusable.

Contact Antigravity Support… zero feedback.

Claude Code: making strides, more community driven, better out-of-box experience, and seems to taking the lead at the moment.

If you want evidence, you can search for this link in the Internet Archive:
https://antigravity.google/assets/docs/plans/plans.md

It previously showed (around February 12) “no weekly rate limits”, but later it was changed to “higher weekly rate limits”.

1 Like

The sheer audacity is absolutely staggering.

Im switching back to claude max, Opus is the only model i want to use, I was enjoying the other google account perks but honestly a cheap pro plan covers all that and i can get all the opus i want in claude cide with the max plan for less money than ai ultra. Im officially done promoting googles ai products it just makes me look like an idiot and a shill

4 Likes

This actually happened, I just used a few minutes to troubleshoot the code after waiting 3 hours due to the limitation, you can see the limitation resource is already 60% left.

Currently I am using the highest package, which is Ultra, but after the last update, Google really lowered the limit and this is very unfair with what I pay every month which is so expensive.

1 Like