Hello Google AI Studio Team,
I am experiencing a persistent issue with the Google AI Studio Build Agent.
The Build Agent repeatedly fails with:
“Quota exceeded. Please try again later.”
I also receive:
“Encountered retryable error from model provider:
Agent execution terminated due to error.”
Project details:
App ID:
020025a0-5b03-41d1-8cb7-702a8bd565b0
Project ID:
gen-lang-client-0046588…
Tier:
Free
The Rate Limit page shows:
Antigravity RPM: 0 / 60
Antigravity TPM: 0 / 100K
Antigravity RPD: 0 / 100
There is currently no billing configured for this project.
I have also tried retrying the Build Agent, but the same error keeps occurring.
The application itself is working, but the Build Agent cannot execute requests.
Please investigate whether there is a separate AI Studio Build/Agent quota, execution limit, provider-routing issue, account/project entitlement issue, or backend issue that is not visible in the Gemini API Rate Limit dashboard.
Could you please check the Build Agent status/entitlement for my App ID and Project ID and let me know whether it can be reset or restored?
I can provide screenshots and exact timestamps if required.
Thank you.
Potential workaround / observation
I was experiencing the same Quota exceeded / resource_exhausted behaviour during AI Studio Build agent tasks. Longer agent runs would work for several minutes (I observed one exceed 300 seconds) and then fail with the quota error.
My project is on a paid tier with an active positive prepaid balance. Normal Gemini requests were working, and the AI Studio Rate Limits page was not showing an exhausted quota.
I started a fresh Build chat within the same existing project/workspace — I did not create a new project or change billing/API keys.
I then asked the new chat to perform a read-only inspection of the existing workspace.
Result:
- First workspace inspection completed essentially immediately.
- Second, substantially larger workspace inspection completed successfully in 67 seconds.
- No quota error occurred.
- The same existing application files/workspace were accessible and intact.
This suggests that, at least in my case, the problem may be related to accumulated Build conversation/context/tool history rather than the published account/project quota itself.
Starting a fresh chat appears to have cleared the problem while retaining the existing workspace.
I can’t confirm whether the underlying cause is context-window size, TPM caused by accumulated context, or something internal to Build’s agent orchestration, but a fresh Build chat may be worth trying for others experiencing this.
It would be helpful if Google could clarify this behaviour. If accumulated Build history can cause resource_exhausted, the UI probably shouldn’t report it simply as “Quota exceeded”, because that strongly suggests the user’s published API/billing quota has been exhausted.
Hello Google AI Studio Team,
I am experiencing the same persistent Google AI Studio Build Agent issue.
My Build Agent edits the project files successfully and the Android project builds successfully, but after a long agent run it ends with:
“Quota exceeded. Please try again later.”
My current observations:
- Model: Gemini 3.8 Flash
- AI Studio App ID: 05866a6c-e89e-4586-a20d-daf9c08cf418
- Antigravity RPM: 0 / 60
- Antigravity TPM: 0 / 100K
- Gemini API Usage: 4 requests
- API success rate: 100%
- API Errors: No data available
- Google AI subscription: Pro
- Android build itself completes successfully before the quota error appears.
Examples of failed Build Agent runs:
- ~651 seconds
- ~1261 seconds
The application preview continues to work normally, and the generated code changes can be built and tested.
Could you please check whether there is a separate Build/Agent execution quota, entitlement, provider-routing issue, or backend restriction that is not reflected in the normal Gemini API Usage/Rate Limit dashboard?
Please check the Build Agent entitlement/quota for my App ID and account.
Thank you.
