Keep getting hit with persistent 503s on antigravity 2.0.1 on Windows. The backend is throwing a capacity exhausted error on cloudcode-pa.googleapis.com and it happens again and again.
I am connecting from Pakistan and wondering if this is a regional capacity squeeze or if my accounts are getting silently shadow banned. Even the browser subagent spawned to report this bug got killed by the exact same 503 capacity error on its first thinking step.
Here is the traceback from the console.
Trajectory ID 45d087cc-ec37-4bf4
Error HTTP 503 Service Unavailable
Sherlog TraceID is 0x992f54c
{
“error”: {
“code”: 503,
“details”: [
{
“@typetype”: “type.googleapis.com/google.rpc.ErrorInfo”,
“domain”: “cloudcode-pa.googleapis.com”,
“metadata”: {
“model”: “gemini-3-flash-agent”
},
“reason”: “MODEL_CAPACITY_EXHAUSTED”
},@type{
“@type”: “type.googleapis.com/google.rpc.RetryInfo”,
“retryDelay”: “15s”
}
],
“message”: “No capacity available for model gemini-3-flash-agent on the server”,
“status”: “UNAVAILABLE”
}
}
EDIT. Tried this across three different accounts now. A free tier, an AI Pro tier, and an AI Ultra tier. All of them throw the exact same 503 capacity exhausted error repeatedly. This rules out any billing/quota limit on a single account. The gateway is completely broken regardless of tier.