Hello Google AI Studio team and community,
I am experiencing persistent failures in Google AI Studio Build that are preventing me from continuing development on my applications. The problem affects multiple apps, not just one project.
1. Description of the problem
AI Studio Build repeatedly produces the following error messages:
-
“Quota exceeded. Please try again later.”
-
“There was an unexpected error. Finish what you were doing.”
-
“An internal error occurred.”
Runs are cancelled unexpectedly, and the coding assistant does not reliably complete the requested work.
The problem has persisted across multiple applications, fresh conversations, different prompts, and different model selections.
Even this minimal test prompt fails:
“Reply with only TEST OK. Do not inspect files, use tools, or modify anything.”
This suggests that the problem is not limited to a particular coding task, although I cannot determine the underlying cause.
2. Evidence from an actual failed session
I have captured a screenshot from AI Studio Build showing these failures in the same development session.
The screenshot records:
-
Model: Gemini 3.8 Flash
-
Run duration: 21 minutes and 6 seconds
-
File edited:
phase32_2PersistentStateAndAdminOps.test.ts -
First visible error: “Quota exceeded. Please try again later.”
-
Additional message: “There was an unexpected error. Finish what you were doing.”
-
Subsequent model attempt: Gemini 3 Flash Preview — Cancelled
-
Final error: “An internal error occurred.”
The failure occurred after the model had already edited a file. This raises an additional concern about interrupted development tasks and whether changes are consistently completed or left partially applied.
I can attach the screenshot as supporting evidence.
3. Troubleshooting already attempted
I have attempted the following:
-
Tested multiple applications in AI Studio Build.
-
Started fresh conversations.
-
Tried different prompts and model selections.
-
Submitted a minimal prompt that did not request file inspection, tool use, or code modifications.
-
Checked the model rate-limit dashboard.
-
Reviewed the billing dashboard and available credit balance.
The rate-limit dashboard for a previously selected Google Cloud project showed zero usage for Gemini 3.8 Flash against the displayed limits. However, I still need to verify that I was viewing the correct project and whether AI Studio Build has separate quotas or restrictions.
The billing dashboard also showed a remaining credit balance. Therefore, I cannot establish that insufficient billing credit is the cause.
4. Impact on development
I am developing Exameno CBE Content Generator, an educational application used for curriculum-aligned teaching and assessment workflows.
I am currently working on important persistence, administration, security, and production-readiness remediation. The repeated failures prevent me from reliably inspecting the codebase, implementing fixes, running follow-up tasks, and completing the required verification.
This is disrupting development across multiple applications and is not merely an isolated failed prompt.
5. Assistance requested from Google
Could the Google AI Studio team please investigate the following?
-
Quota enforcement: Why does AI Studio Build return “Quota exceeded” when the displayed model rate-limit dashboard appears to show available capacity? Could there be separate Build-mode, project-level, account-level, or tool-execution quotas?
-
Internal errors: Is there a known issue causing “An internal error occurred” and unexpected cancellations in AI Studio Build?
-
Multiple applications: Could an account-level restriction, project configuration issue, or service-side problem explain why multiple applications are affected?
-
Model availability: Is there a known problem affecting Gemini 3.8 Flash or Gemini 3 Flash Preview in AI Studio Build?
-
Interrupted changes: What is the recommended procedure for verifying and recovering changes when a run fails after editing files?
-
Diagnostics: Which logs, request identifiers, timestamps, or other diagnostic information should I provide to help identify the root cause?
-
Resolution: What steps can I take to restore reliable operation without risking existing application code or workspace changes?
6. Expected outcome
I would appreciate confirmation of whether this is a known service issue, a quota or account restriction, a project configuration problem, or another underlying failure.
If further information is needed, I can provide screenshots, approximate timestamps, model names, and relevant project configuration details. I will not share API keys, service-account credentials, passwords, or other secrets publicly.
Thank you for your assistance. I would appreciate guidance from the Google AI Studio team or from anyone in the community who has experienced and successfully resolved the same problem.
