Antigravity reset · community

Community reports

First-hand experiences, with dates, context, and the limits of what they tell us.

Reviewed

User report · Quota

A largely unused Pro account showing 0% Argon quota

The author reports that a second paid Pro account showed no weekly Argon quota despite very little prior Antigravity use. Restarting, updating, and testing a fresh local profile did not change the displayed balance.

Context Windows desktop · self-reported paid AI Pro · two accounts on the same computer and network · client version not stated

What this does not prove

This does not establish a shared device quota, a failed scheduled reset, or a confirmed backend fault. Account eligibility and actual provisioning were not independently verified. Switching accounts is not a demonstrated remedy.

At review, the thread had no replies or confirmed resolution. A visible countdown alone does not explain why the balance is zero.

Carlos_Morales_avi · Google AI Developers ForumRead original ↗
User report · Quota

Argon and Flash: one user’s quota comparison

The author compared Argon High with Flash 3.8 High and reported different quota consumption relative to published API prices. They also observed that usage readings could settle several minutes after a task.

Context Antigravity 2.27.0 · Ultra 5× · Windows · one run per model

What this does not prove

This is a small, self-reported comparison. API prices do not establish a subscription quota conversion rate; the reported ratio is not an official multiplier.

Later replies add tests on other surfaces. We have not independently reproduced the result.

Yiji_Kong · Google AI Developers ForumRead original ↗
User report · Performance

Desktop slowdown during long agent sessions

The author reported animation-related CPU activity and a language-server process reaching about 2.9 GB after hours of use. Long conversations also felt increasingly heavy.

Context Antigravity App 2.26.0 · Windows 11 · Ryzen 7 5800H · 32 GB RAM

What this does not prove

The author could not distinguish a memory leak from an unbounded cache. Desktop UI performance is separate from model reasoning quality.

A later reply suggests reduced motion and version 2.28.0. We have not verified a fix; the memory concern remains unconfirmed.

Iamemily2050 · Google AI Developers ForumRead original ↗
User report · Experience

A positive first impression of Argon for coding

The author described a responsive experience with multi-file TypeScript refactoring and new modules, and reported that context held up well during their work.

Context Antigravity App · Argon Medium · personal coding workflow

What this does not prove

A first impression, not a controlled benchmark. The author also promotes a personal development tool in the post.

No independent follow-up test is included in this collection yet.

Roedy_Rustam · Google AI Developers ForumRead original ↗