My Google AI Studio Build project has been unable to generate for approximately 48 hours.
The project has billing enabled and is currently Tier 1 Prepay. The project has incurred approximately $0.76 in Gemini API usage.
AI Studio repeatedly returns:
“You are out of generations for this project. Please choose a different project.”
The agent also returns:
“Encountered retryable error from model provider: Agent execution terminated due to error.”
and:
“Quota limits exceeded”
The important part is that the normal Gemini API quotas are NOT exhausted.
Current Gemini API rate limits shown in AI Studio:
Gemini 3.7 Flash:
RPM: 16 / 1,000
TPM: 1.28M / 2M
RPD: 69 / 10K
Antigravity:
RPM: 0 / 30
TPM: 0 / 200K
RPD: 0 / 1K
Deep Research Pro Preview:
RPM: 0 / 1
TPM: 0 / 500K
RPD: 0 / 1.44K
Other models are also showing zero or very low usage.
I have not actively used AI Studio for several days, yet this project continues to report that it is out of generations.
I have also tested multiple API keys. The exact same error occurs regardless of which API key I select, including API keys associated with other AI Studio projects that work normally.
Other AI Studio projects on the same Google account are working correctly.
This therefore does not appear to be a normal Gemini API rate limit, an API key problem, or an account wide billing problem.
The project was previously working normally. I was using Gemini 3.7 Flash in AI Studio Build Mode to develop a Python/Django website. The agent successfully created and modified the Django application, ran commands, installed dependencies, tested the application, and successfully completed:
python3 manage.py check
Result:
System check identified no issues (0 silenced).
It also successfully ran:
python3 manage.py collectstatic --noinput
Result:
130 static files copied.
The Django application is currently healthy and has been committed to GitHub.
The failure began while the AI Studio Build agent was troubleshooting its development preview environment. The agent encountered repeated container and dev server failures involving server.ts, port 3000, the AI Studio control plane, and Vite middleware. The project itself was not exceeding its visible Gemini API quotas.
The AI Studio container subsequently timed out and was reprovisioned. The Django source survived and the agent was able to recover the environment. However, AI Studio Build generation continued to fail with the quota error.
I am requesting that Google investigate the AI Studio Build generation state or backend quota/entitlement state for this specific project.
I have attached screenshots showing:
- The project and its Tier 1 Prepay billing status.
- Gemini API usage for the project.
- Gemini API rate limits and current usage.
- The AI Studio “You are out of generations for this project” error.
- The repeated agent execution and quota errors.
This has persisted for approximately 48 hours despite billing being enabled, Tier 1 status being active, visible quotas being available, and other AI Studio projects working normally. I have also tried generating an API key with a Google Cloud account with payment, no luck.
I have over 25$ in my account that CANNOT be used at this point. I waited days to make sure it was a daily limiter.
Please investigate whether the generation entitlement or AI Studio Build agent state for this project is stuck or incorrectly reporting that the project has exhausted its generations.