Dear Antigravity Support Team,
I am a Google AI Ultra subscriber (confirmed as “the best plan” in the error message) and primarily using Claude Opus 4.6 in Antigravity IDE.
I am facing a confusing and inconsistent quota display issue that is blocking my workflow:
-
In Settings → Models (quota section), it clearly shows ~20% tokens/credits remaining for Opus 4.6.
-
However, when using the chat/agent, I get this message: “Your plan’s baseline quota will refresh on 3/23/2026, 6:29:09 PM. You are subscribed to the best plan.”
According to the official docs, Ultra plan should have:
-
The highest generous quota
-
Refreshed every 5 hours (rolling window)
-
No weekly baseline rate limit / cap
This seems like a bug because:
-
A “baseline quota” with a specific future refresh time (several hours away) should not appear on Ultra, especially when settings still shows 20% available.
-
I have been using Opus 4.6 heavily today on large codebases (refactoring big projects, generating features, multiple complex agent sessions), but I have not exhausted my quota — I always keep a buffer and never hit zero before.
Additional details:
-
Antigravity version: 1.20.6 (latest as of March 2026)
-
OS: macOS Sonoma 14.6.1
-
AI Credits toggle: Off (set to “Never” – do not use overages)
-
Approximate usage today: Heavy but reasonable use on large codebases — several agent tasks, no extreme/spam usage. Still far from any reasonable limit.
-
Steps tried to fix: Signed out and back in, restarted IDE multiple times — issue persists.
Questions:
-
Is this a known display/sync bug in v1.20.6 (similar to reports in forums about quota mismatch after AI Credits update)?
-
Why does a baseline refresh timer appear on Ultra plan, when docs state no weekly limit and only 5-hour rolling refresh?
-
Can you manually sync/reset my quota entitlement or confirm if it’s a server-side sync issue with Ultra subscription?
-
How to regain immediate access to Opus 4.6 without waiting until 6:29 PM?
This is significantly impacting my productivity as an Ultra user expecting generous, uninterrupted access.
Thank you for your prompt help — looking forward to your response within 24 hours as per support SLA.
