Hello,
I am using Antigravity (v2.7.3) with a Google One AI Pro account (ed**@gmail.com) and noticed an inconsistency between the UI usage panel and the actual API behavior.
Issue:
The Antigravity Panel displays 100% available for both Gemini and Claude, with “4h 59m” remaining and “0.0 pp consumed” over the last 7 days. Despite this, agent execution fails with the following error:
Trajectory ID: 7feaa790-a673-4343-be9c-***
TraceID: 0x5cfa7d5*
HTTP 429 Too Many Requests
{
“error”: {
“code”: 429,
“message”: “Resource has been exhausted (e.g. check quota).”,
“status”: “RESOURCE_EXHAUSTED”
}
}
Relevant headers:
X-Cloudaicompanion-Trace-Id: 5cfa7d5**
Server: ESF
Server-Timing: gfet4t7; dur=468
Date: Fri, 18 Sep 2026 13:08:31 GMT
What I expected:
If the panel shows 100% quota available, agent execution should work normally, or the panel should accurately reflect the actual quota status causing the block.
What happened:
The agent terminates with an error (“Agent terminated due to error”) with no visual indication on the dashboard that the actual quota (apparently measured at another layer, via Cloud AI Companion / Vertex AI) was exhausted.
Additional note:
When exporting the session logs (via diagnostic export), all 24 log files (.log) came back empty (0 bytes), including Antigravity.log, network.log, main.log, and cloudcode.log. This made troubleshooting on my own difficult and may be a separate bug in the log export mechanism.
Environment information:
- Antigravity v2.7.3
- Account: ed****@gmail.com (Google One AI Pro)
- Model selected at the time of the error: Gemini 3.1 Pro (High)
- Date/time of error: 09/18/2026, 13:08:31 G