Provider pricing hub
Gemini API Pricing (July 2026)
The Gemini API costs $1.50/$9 per million input/output tokens for Gemini 3.5 Flash and $2/$12 for Gemini 3.1 Pro (rising to $4/$18 above 200K tokens). The older Gemini 3 Flash remains available at $0.50/$3, and Gemini 3.1 Flash-Lite is the budget floor at $0.25/$1.50. Every current tier ships a 1M-token context window.
Google prices the Gemini API in three lanes: Pro for frontier reasoning ($2/$12), Flash for the price-performance middle ($1.50/$9 on 3.5 Flash), and Flash-Litefor high-volume work ($0.25/$1.50) — with a long-context surcharge on Pro prompts past 200K tokens. The table below lists every Google model in BenchLM's pricing registry, synced from Google's Gemini Developer API pricing page.
Whether the rates are worth paying is a benchmark question, not a pricing one: see the Gemini model rankings and the price-vs-performance view for how each tier scores per dollar.
Every Gemini API price per 1M tokens
| Model | Input $/M | Cached input $/M | Output $/M | Context |
|---|---|---|---|---|
| Gemini 3 Pro | $2 | $0.2 | $12 | 2M |
| Gemini 3.1 Pro | $2 | $0.2 | $12 | 1M |
| Gemini 3.5 Flash | $1.5 | $0.15 | $9 | 1M |
| Gemini 2.5 Pro | $1.25 | $0.125 | $10 | 1M |
| Gemini 1.5 Pro | $1.25 | $0.125 | $5 | 1M |
| Gemini 3 Flash | $0.5 | $0.05 | $3 | 1M |
| Gemini 2.5 Flash | $0.3 | $0.03 | $2.5 | 1M |
| Gemini 3.1 Flash-Lite | $0.25 | $0.025 | $1.5 | 1M |
Standard-mode rates for prompts up to 200K tokens; Google's published long-context surcharges (e.g. Gemini 3.1 Pro at $4/$18 above 200K) are not included. Cached input is a 10% cache-read estimate except Gemini 3.5 Flash, whose $0.15/M cache-hit rate is documented. Google lists lower Batch pricing for some tiers but no flat lineup-wide discount, so batch columns are omitted.
Estimate your monthly Gemini API bill
| Model | Est. monthly cost |
|---|---|
| Gemini 3 Pro | $44 |
| Gemini 3.1 Pro | $44 |
| Gemini 3.5 Flash | $33 |
| Gemini 2.5 Pro | $32.5 |
| Gemini 1.5 Pro | $22.5 |
| Gemini 3 Flash | $11 |
| Gemini 2.5 Flash | $8 |
| Gemini 3.1 Flash-Lite | $5.5 |
Estimates use Gemini's standard per-token rates from the table above. For task-level presets and cross-provider comparison, use the full AI cost calculator.
How Google’s tiering works: Pro, Flash, Flash-Lite — and the 200K cliff
Every Gemini generation ships the same three-tier ladder. Pro is the frontier tier: Gemini 3.1 Pro and Gemini 3 Pro both list at $2/$12 per million tokens. Flash is the workhorse: Gemini 3.5 Flash at $1.50/$9 has drifted up-market — it now costs more than the 2.5-era Pro ($1.25/$10 input-side) and supersedes Gemini 3 Flash ($0.50/$3), which stays on the price sheet. Flash-Lite ($0.25/$1.50 for Gemini 3.1 Flash-Lite) is the volume floor.
The detail that surprises budgets is long-context pricing. Google's pricing page lists the Pro rates for prompts up to 200K tokens, rising to $4/$18 above that on Gemini 3.1 Pro and 3 Pro; Gemini 2.5 Pro jumps from $1.25/$10 to $2.50/$15 past 200K, and the legacy 1.5 Pro doubles past 128K. If you routinely stuff the 1M-token window, model the doubled rate — the estimator above uses standard-tier prices.
Gemini vs OpenAI and Claude pricing
At the flagship tier Google undercuts both rivals on paper: Gemini 3.1 Pro at $2/$12 lists below GPT-5.6 Sol ($5/$30) and Claude Opus 4.8 ($5/$25) — under half the price on both input and output. The tighter fight is the mid-tier: Gemini 3.5 Flash ($1.50/$9) sits between GPT-5.6 Terra ($2.50/$15) and GPT-5.6 Luna ($1/$6), and below Claude Sonnet 5 ($3/$15). At the floor, Flash-Lite ($0.25/$1.50) beats Haiku 4.5 ($1/$5) by 3–4x.
The caveat cuts both ways: rivals' discounts are more legible — Anthropic publishes a flat 50% batch rate and ~10% cache reads, while Google's batch and long-context math varies by tier. Compare the full tables at OpenAI API pricing and Claude API pricing, or run head-to-heads on the compare hub.
Which Gemini tier should you pick?
Start with Gemini 3.5 Flash($1.50/$9) as the default: it scores 75 on BenchLM's leaderboard — even with Gemini 3 Pro — at a fraction of frontier pricing, and it is the tier Google itself pushes hardest. Escalate to Gemini 3.1 Pro ($2/$12, scoring 89) for hard reasoning, agentic coding, and work where quality failures cost more than tokens; the upgrade is only ~33% more per input token as long as you stay under the 200K surcharge line.
Route the boring volume — classification, extraction, tagging, short summaries — to Gemini 3.1 Flash-Lite ($0.25/$1.50), which is 6x cheaper than 3.5 Flash on both sides of the meter. The previous-generation Gemini 3 Flash ($0.50/$3) and 2.5 Flash ($0.30/$2.50) remain priced if you have pinned workloads, but for new builds the 3.1/3.5 lineup dominates. A three-way routing split typically lands the blended rate well below running everything on one tier — model yours with the cost calculator.
Gemini API pricing FAQ
How much does the Gemini API cost?
Gemini API pricing runs from $0.25/$1.50 per million input/output tokens (Gemini 3.1 Flash-Lite) through $0.50/$3 (Gemini 3 Flash) and $1.50/$9 (Gemini 3.5 Flash) to $2/$12 for Gemini 3.1 Pro — with Pro rates rising to $4/$18 on prompts above 200K tokens. The full table above lists every current and legacy Google model.
Is the Gemini API free?
Google offers a rate-limited free tier of the Gemini API through AI Studio, which is enough for prototyping but not production traffic. Paid usage bills per million tokens at the rates above — the cheapest sustained path is Flash-Lite at $0.25/$1.50. Google's open-weight Gemma models can also be self-hosted at no per-token cost.
What is Google Gemini API pricing per 1M tokens?
Google bills the Gemini API per million tokens with separate input and output rates: Gemini 3.1 Pro at $2 input / $12 output, Gemini 3.5 Flash at $1.50/$9, Gemini 3 Flash at $0.50/$3, and Gemini 3.1 Flash-Lite at $0.25/$1.50. Pro tiers carry a documented long-context surcharge above 200K tokens ($4/$18 on 3.1 Pro).
How does Gemini pricing compare to GPT and Claude?
Gemini 3.1 Pro ($2/$12) lists at less than half of GPT-5.6 Sol ($5/$30) and Claude Opus 4.8 ($5/$25), and Flash-Lite undercuts every OpenAI and Anthropic budget tier. Mid-tier, Gemini 3.5 Flash ($1.50/$9) sits between GPT-5.6 Luna and Terra. See the OpenAI pricing hub and Claude pricing hub for the full rival tables.