paid tier / month
per day
per 1k requests
vs GPT-4o / mo

Gemini vs the field — same workload

Identical tokens and volume, priced on each model.

ModelInput $/1MOutput $/1MCost / month
⚠️ Gemini reference pricing (June 2026): 2.5 Pro $1.25/$10, 2.5 Flash $0.30/$2.50, 2.0 Flash $0.10/$0.40 per 1M tokens. The free tier (AI Studio) has rate limits, not token charges. Confirm on Google's pricing page. See every provider on the full AI API cost calculator.

The free tier comes first

Gemini is unusual among the big models: it has a genuine free tier through Google AI Studio, with no credit card to start. It's capped by requests-per-minute and requests-per-day rather than tokens, so for prototypes and many low-traffic apps the honest answer is you may pay nothing. This calculator shows the paid-tier bill — what you'd owe once you exceed the free limits and switch to pay-as-you-go.

When you do pay, Gemini is one of the cheapest capable options. 2.0 Flash ($0.10 / $0.40) is astonishingly cheap for high-volume work, 2.5 Flash ($0.30 / $2.50) adds capability, and 2.5 Pro ($1.25 / $10) competes with frontier models at a lower input price. As always, output is the costly side — cap answer length first. Comparing providers? Put Gemini against GPT-4o and Claude on the AI API cost calculator, price a feature with the AI app cost estimator, and get your key via the Gemini guide.

FAQ

Is Gemini really free? The free tier is real but rate-limited; high volume needs the paid tier shown above.

Which Gemini is cheapest? Gemini 2.0 Flash, at about $0.10 input / $0.40 output per 1M tokens.