Understanding the Gemini API pricing structure and rate limit quotas is critical for building cost-effective production applications. Google offers both a Free Tier in Google AI Studio for prototyping and a Pay-As-You-Go Tier for production workloads.
Free Tier vs Pay-As-You-Go Model Overview
| Feature | Free Tier (AI Studio) | Pay-As-You-Go Tier |
|---|---|---|
| Price | $0.00 | Pay per 1M tokens |
| Data Privacy | Prompts used to train models | Prompts NOT used to train models |
| Gemini 1.5 Flash RPM | 15 RPM | 1,000 RPM |
| Gemini 1.5 Pro RPM | 2 RPM | 360 RPM |
| Max Context Window | 1,000,000 Tokens | 2,000,000 Tokens |
Token Pricing Breakdown (Pay-As-You-Go)
Google charges based on input tokens (prompt text/images/audio) and output tokens generated by the model.
1. Gemini 1.5 & 2.0 Flash (High Speed / Low Cost)
- Input Tokens (prompts < 128k): $0.075 per 1 million tokens
- Input Tokens (prompts > 128k): $0.15 per 1 million tokens
- Output Tokens: $0.30 per 1 million tokens
- Context Caching: $0.01875 per 1M tokens/hour
2. Gemini 1.5 Pro (Deep Reasoning / Complex Code)
- Input Tokens (prompts < 128k): $1.25 per 1 million tokens
- Input Tokens (prompts > 128k): $2.50 per 1 million tokens
- Output Tokens: $5.00 per 1 million tokens
Understanding Gemini API Rate Limits
Rate limits are measured across three metrics:
- RPM (Requests Per Minute): Number of individual API calls per 60 seconds.
- TPM (Tokens Per Minute): Total combined input + output tokens processed per minute.
- RPD (Requests Per Day): Hard cap on daily API invocations (1,500 RPD on Free Tier).
429 Quota Exceeded -> Solution: Implement exponential backoff or upgrade to Pay-As-You-Go billing.FAQ
Does Google train on Free Tier API data?
Yes. On the Free Tier in Google AI Studio, input prompts and output responses may be reviewed by human annotators to improve Google models. If handling sensitive PII data, use the Pay-As-You-Go Tier.
How do I increase my Gemini API rate limits?
To increase RPM and TPM limits beyond the free quotas, add a credit card in Google Cloud Console under Billing -> Linked AI Studio Project.
