Hello Google Antigravity Support Team,
For the past few days, I have repeatedly encountered “Agent execution terminated due to error” when submitting requests in Antigravity. This is preventing me from working on my projects.
My account currently shows “Available AI Credits: 35000,” and “Enable AI Credit Overages” is switched on. Despite this, requests fail with HTTP 429 and RESOURCE_EXHAUSTED.
The attached screenshots show the failed request with Claude Sonnet 4.6 (Thinking) selected, along with my available credits and enabled credit overages.
The latest “Copy debug info” contains:
Trajectory ID: 8a38160d-cc25-46cf-940c***
TraceID: 0xe277484d*
Error: HTTP 429 Too Many Requests
Timestamp: 20 September 2026, 04:33:58 GMT
X-Cloudaicompanion-Trace-Id: e277484d3**
Error response:
{
“error”: {
“code”: 429,
“message”: “Resource has been exhausted (e.g. check quota).”,
“status”: “RESOURCE_EXHAUSTED”
}
}
Please investigate why requests are being rejected while I have 35,000 available AI credits and credit overages enabled. If a separate rate limit, model quota, or service capacity restriction applies, please identify which restriction is affecting my account and explain when it resets.
Please also confirm whether my available credits can be used for the selected model and provide the steps needed to restore access.
A clearer error message explaining the specific restriction and expected recovery time would be very helpful.
Thank you for your assistance.
Best regards,
José María Cabeza Rodríguez
Open Google Antigravity and a local project.
- Select “Claude Sonnet 4.6 (Thinking)” and Plan mode.
- Confirm that “Enable AI Credit Overages” is enabled and the account shows “Available AI Credits: 35000.”
- Submit a prompt asking the agent to investigate a project issue.
- The agent stops with “Agent execution terminated due to error.”
- Click “Copy debug info.” The output shows HTTP 429 Too Many Requests, status RESOURCE_EXHAUSTED, and “Resource has been exhausted (e.g. check quota).”
Expected result: The agent processes the request, using available AI credits where applicable.
Actual result: The request fails despite 35,000 available AI credits and credit overages being enabled. This issue has occurred repeatedly over the past few days.
