What causes "Gemini API Error 429 Resource Exhausted"?

Per-minute request or token limit (RPM/TPM) exceeded Daily request quota (RPD) used up; it resets at midnight Pacific time Free tier limits are far lower than paid tiers and differ per model Quota is per project, so several keys in one project share it Temporary service-side capacity problems that return 429 below your limits

How do I fix "Gemini API Error 429 Resource Exhausted"?

1. Check which limit you hit: Open Google AI Studio and go to the usage / rate limit page for your project Compare requests per minute, tokens per minute and requests per day against your model's limits The error message lists the quota metric that was exceeded 2. Retry with exponential backoff: Wait for the retryDelay shown in the error, then retry Use exponential backoff with jitter instead of immediate retries Queue or batch requests so bursts stay under RPM 3. Enable billing or move up a tier: In AI Studio open the API key page and click Set up billing for the project Paid tiers raise RPM and RPD limits; request a quota increase if you still hit the ceiling 4. Switch to a lighter model or wait for reset: Use a Flash or Flash-Lite model, which have higher limits than Pro If the daily quota is gone, wait until midnight Pacific time 5. Rule out an outage: Search the Google AI Developers forum for same-day 429 reports If you are well below limits on a paid tier, post your project ID and model on the forum

Gemini API Error 429 Resource Exhausted

Workaround available API, Google AI Studio Updated November 14, 2026Based on 6 user reportsHow we research

Quick answer

A 429 RESOURCE_EXHAUSTED from the Gemini API means you hit a per-minute rate limit or a daily quota for that model and project. Back off and retry, check your usage in AI Studio, or enable billing; some paid Tier 1 users also report 429s well below limits during incidents.

Most likely fix: Check which limit you hit

How to fix it

Try these in order — the most common fix is first.

1. Check which limit you hit
1. Open Google AI Studio and go to the usage / rate limit page for your project
2. Compare requests per minute, tokens per minute and requests per day against your model's limits
3. The error message lists the quota metric that was exceeded
Did this fix it?
2. Retry with exponential backoff
1. Wait for the retryDelay shown in the error, then retry
2. Use exponential backoff with jitter instead of immediate retries
3. Queue or batch requests so bursts stay under RPM
Did this fix it?
3. Enable billing or move up a tier
1. In AI Studio open the API key page and click Set up billing for the project
2. Paid tiers raise RPM and RPD limits; request a quota increase if you still hit the ceiling
Did this fix it?
4. Switch to a lighter model or wait for reset
1. Use a Flash or Flash-Lite model, which have higher limits than Pro
2. If the daily quota is gone, wait until midnight Pacific time
Did this fix it?
5. Rule out an outage
1. Search the Google AI Developers forum for same-day 429 reports
2. If you are well below limits on a paid tier, post your project ID and model on the forum
Did this fix it?

Symptoms

API calls return HTTP 429 with 'Resource has been exhausted (e.g. check quota)' or 'You exceeded your current quota', often with a retryDelay value, even on low traffic.