Genace
Sign inGet started

You exceeded your current quota, please check your plan and billing details. — Veo 3.1 API 429

HTTP 429 RESOURCE_EXHAUSTED Retry depends on cause

The exact error 

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.

A Veo request through the Gemini API (an API key, not a Vertex project) was rejected with 429 RESOURCE_EXHAUSTED and the Gemini API’s standard quota message. For Veo 3.1 the cause is usually a very small daily request allowance, but users also report this error on paid projects that show zero usage.

What it means

This is the Gemini API’s generic “quota exceeded” response. A Veo developer posted it word for word in forum thread 97560 while calling veo-3.0-generate-preview with client.models.generate_videos. A forum responder there said Veo access on the free tier is very limited. The Veo 3.1 threads below report the same 429 RESOURCE_EXHAUSTED status; they do not all quote the message text.

What Google documents on its rate limits page:

  • Limits are applied per project, not per API key, and requests-per-day (RPD) quotas reset at midnight Pacific time.
  • “Rate limits are more restricted for experimental and preview models.” Google’s Veo docs list Veo 3.1, Veo 3.1 Fast and Veo 3.1 Lite as Preview.
  • Separate spend-based limits apply on a rolling 10-minute window by usage tier (the page lists $10 per 10 minutes for Tier 1 at the time of writing), and hitting one also returns 429 RESOURCE_EXHAUSTED.
  • Your actual limits are shown in AI Studio. The rate limits page does not publish Veo numbers.

What users report, not confirmed by Google:

  • Several Tier 1 users say their dashboard shows a Veo limit of 10 requests per day per model (threads 172205, 175417, 175949). Check your own AI Studio limits; do not rely on that number.
  • In thread 175417, a Tier 1 prepay user saw 0/10 daily usage in AI Studio, yet every veo-3.1-lite-generate-preview request failed instantly with 429.
  • In thread 179821, a user who had recently moved to Tier 1 got 429 on every Veo 3.1 Lite and Fast request while Gemini text models worked, AI Studio showed 0 RPM and 0 RPD, the response had no Retry-After, QuotaFailure or ErrorInfo, and it happened across more than one project.
  • In thread 172205, a Tier 1 user saw the first Veo request of the day rejected for more than 48 hours, as if the daily quota never reset.
  • In thread 175949, a user reports that Veo requests blocked by the safety filter still count against the daily allowance.

How to confirm it’s this

  1. Status. HTTP 429, status RESOURCE_EXHAUSTED, on a Veo model, through the Gemini API.
  2. Read the error details. Gemini API quota errors can include a QuotaFailure entry with quotaMetric and quotaId, and a RetryInfo with a retryDelay, as in the example quoted in thread 109580. If the metric names a free_tier quota with limit: 0, you are on the free-tier problem instead (see below).
  3. Compare with AI Studio. Look at the project’s Veo RPD usage. If it is at the limit, this is a real daily cap. If it shows no usage, you are in the situation users describe in threads 175417 and 179821.
  4. Check the project. Make sure the key belongs to the project you upgraded; limits are per project.

Fix

  • Daily cap reached: wait for the reset at midnight Pacific time, and plan batch jobs around a small number of Veo requests per day.
  • Spend limit: the rate limits page says to wait and retry after a short period, reduce the rate of expensive requests, and request a rate limit increase if you hit it during normal use.
  • Filtered attempts: if many attempts end in safety or audio filter results, count those against your daily budget too, based on the report in thread 175949.
  • Tier 1 with zero usage and instant 429s: retrying does not fix this in the reports above. In thread 179821 a forum responder asked for the project ID and billing ID by direct message plus the full JSON error. Collect those and send them privately, never in a public post.

Retry or not

Depends on which limit you hit.

  • A per-minute or spend-based limit is transient. Google’s troubleshooting guide recommends exponential backoff with jitter for 429 and notes the Python SDK already retries up to four times. If the error carries a retryDelay, wait at least that long.
  • A daily cap will not clear until midnight Pacific time. Backing off for seconds or minutes wastes calls.
  • The zero-usage Tier 1 case is not fixed by retrying, according to every report above.

For agents: retry briefly with backoff, then stop and report “Veo quota exhausted” with the quota details you found, instead of looping for hours.

When it’s not this

  • Resource exhausted, please try again later. The Vertex AI version: shared pay-as-you-go capacity on a Google Cloud project, not a Gemini API key.
  • generate_content_free_tier_requests, limit: 0. A free-tier metric with no allowance. Upgrading the project’s tier is the fix, not waiting.
  • The model is overloaded. A 503 capacity problem on Google’s side, not your quota.

Error text last verified against the sources below on .

Where these facts come from