We are getting an error calling the Gemini API : Error calling model 'gemini-3.6-flash' (RESOURCE_EXHAUSTED

Google AI Studio
,
1

we are getting an error calling the Gemini API : Error calling model ‘gemini-3.6-flash’ (RESOURCE_EXHAUSTED): 429 RESOURCE_EXHAUSTED. {‘error’: {‘code’: 429, ‘message’: 'You exceeded your current quota, please check your plan and billing details. For more information on this error, head to: https://ai.google.dev/gemini-api/docs/rate-limits.

2

Hi @Nitin_Vilas_Kadam,
We recently made some backend updates to ensure that all 429 errors come with a description field which properly explains the cause. We recommend scanning your logs for these. While a lot of pertinent information is currently available in the AI Studio Usage, Rate Limit & Billing pages, we’re working to make the cause of 429 errors easier to find in AI Studio. Please see below for advice on specific categories of 429 errors.

Tier drop

If your account has recently experienced a tier drop (e.g. tier 3→2 or 2→1). This is very likely the root cause of the 429’s as lower tiers come with lower rate limits & spend caps.

This happens if an automated flag is applied to your account. Unfortunately our support team can’t check or remove these flags. Please check for banners / error callouts in the AI Studio Projects page & the Billing page. You should also check whether you’ve received an email from us with an appeal link.

Prepay funds depleted

If your prepay balance is depleted, we’re unable to serve responses. Generally this can be fixed by buying credits. Note that GCP Free Trial credits can’t be used to pay for Gemini API usage (details).

Spend cap exceeded

We return 429 errors if a spend cap is exceeded.

Rate limits

Multiple different types of rate limits can result in 429 errors. You should check the AI Studio Rate Limits page to see details of these. Notes:

  • Be sure to select the applicable project & time range.

  • Take special care to check the Search grounding & Map grounding limits further down the page

  • If you’re running into limitations from RPM or TPM limits, consider implementing backoff and retry with a long maximum delay (our SDKs support this through retry config).