Google Gemini API API pricing
Long-context multimodal models with the most usable real free tier via AI Studio, backed by Google Cloud scale.
Free tier or trialBilled per token·Verified Jun 27, 2026
Free tier
Free tier on Flash models: ~1,500 req/day, 250k TPM (Pro models paid-only since Apr 2026)
Plans and prices
| Plan | Price | What it covers |
|---|---|---|
| Free tier (AI Studio) | $0 | Free access to Gemini models via Google AI Studio with rate/quota limits; data may be used to improve products. |
| Gemini 2.5 Flash-Lite | $0.10 / $1.50 in/out per 1M | Lowest-cost tier; output $0.40 on the Developer API. Batch halves it to $0.05/$0.20. |
| Gemini 2.5 Flash | $0.30 / $2.50 in/out per 1M | Cost-efficient workhorse with 1M context; Batch $0.15/$1.25; caching $0.03 per 1M + $1.00/hr storage. |
| Gemini 2.5 Pro | $1.25 / $10.00 in/out per 1M | High-reasoning tier; rises to $2.50/$15.00 for prompts over 200K tokens. Batch $0.625/$5.00. |
| Grounding / tools add-on | $35 per 1,000 prompts | Google Search grounding after free tier; Maps grounding $25 per 1,000 prompts. |
What to watch for on the bill
- Restrictive quotas/rate limits on free and lower paid tiers cause friction
- Intelligence Index sits mid-pack rather than top-of-class vs. similarly-priced reasoning models
- Free-tier inputs may be used to improve Google products, a data-privacy concern for some teams
The pricing page, as we saw it
Source: https://ai.google.dev/gemini-api/docs/pricing. Prices change. Confirm on the vendor page before you commit.
Compare Google Gemini API
- OpenAI API vs Google Gemini API
- Anthropic Claude API vs Google Gemini API
- Google Gemini API vs Mistral La Plateforme
Other LLM APIs
OpenAI API · Anthropic Claude API · Mistral La Plateforme · xAI Grok API · Groq · DeepSeek API