Quick answer
Gemini API documents rate-limit dimensions as RPM, TPM, and RPD vary by model and usage tier. Limits apply per project, not per API key, and daily requests reset at midnight Pacific Time.
Verified Aug 22, 2026Official source
Current limits
| Constraint | Current value | Scope | Verified source |
|---|---|---|---|
| rate-limit dimensionsLimits apply per project, not per API key, and daily requests reset at midnight Pacific Time. | RPM, TPM, and RPD vary by model and usage tier | Per Google Cloud project | GoogleAug 22, 2026 |
Why does this limit matter?
Creating additional keys does not multiply project capacity, and any exhausted dimension can reject a request.
This value is scoped to Per Google Cloud project; a different plan, runtime, model, endpoint, region, or account can produce a different effective constraint.
What should you check?
- Open the active project in AI Studio and match the exact model to its displayed tier limits.
- Confirm the exact plan, model, runtime, endpoint, region, and account that serve the failing workload.
- Record the observed value, response headers or configuration, timestamp, and source without logging secrets.
Important caveats
- Preview and experimental models can have more restrictive limits than stable models.
- Treat the official source and live account configuration as authoritative if they differ from this verified snapshot.
HyperObserve reports the documented platform constraint. Your application, SDK, gateway, provider, region, or account can impose a lower effective limit.
Related