Gemini 429 RESOURCE_EXHAUSTED

Retry with exponential backoff

Standard retry pattern with full jitter.

from google.genai import errors
import random, time

def call_with_backoff(fn, max_attempts=5):
    for attempt in range(max_attempts):
        try:
            return fn()
        except errors.APIError as e:
            if e.code != 429 or attempt == max_attempts - 1:
                raise
            wait = random.uniform(0, min(60, 2 ** attempt))
            time.sleep(wait)

Upgrade from free to paid tier

Enable billing in Google AI Studio or GCP to unlock significantly higher quotas (1,000+ RPM vs. 15 RPM).

Request quota increase on Vertex AI

In GCP Console, navigate to IAM & Admin > Quotas, filter for 'Vertex AI', find the specific model quota, and request an increase. Usually approved within 24-48 hours.

Fail over to a cheaper model

Gemini Flash has separate quotas from Pro — failover routing distributes load.

MODELS = ["gemini-3.5-pro", "gemini-3.5-flash"]

def call_with_model_fallback(prompt):
    for model in MODELS:
        try:
            return client.models.generate_content(
                model=model, contents=prompt,
            )
        except errors.APIError as e:
            if e.code == 429:
                continue
            raise