Gemini 3.1 Pro Pricing Explained (2026): Google's Mid-Tier Flagship at $1.25/$10
Gemini 3.1 Pro costs $1.25/$10 per 1M tokens with a 1M context window — tied on price with GPT-5.5, ~41% cheaper blended than Claude Sonnet 5, and a meaningful step up over Gemini 2.5 Pro at the same list price.
Gemini 3.1 Pro costs $1.25 per 1M input tokens and $10 per 1M output tokens, with a 1,000,000-token context window and 65,536-token max output. That's identical per-token pricing to GPT-5.5, ~41% cheaper blended than Claude Sonnet 5 ($3/$15), and a meaningful step up in capability over the older Gemini 2.5 Pro at the same list price. Use the interactive Gemini pricing calculator to model your own workloads.
By the LLMCalculator.net Research Team · Last updated: July 21, 2026
Key facts (verified live)
| Field | Value |
|---|---|
| Provider | |
| Input price | $1.25 / 1M tokens |
| Output price | $10 / 1M tokens |
| Context window | 1,000,000 tokens |
| Max output | 65,536 tokens |
| Knowledge cutoff | February 2026 |
| Speed | Medium |
| Intelligence index | 57 |
| Best for | Reasoning, vision, long-context |
Source: LLMCalculator live pricing API (/api/v1/models), snapshot 2026-07-20. All numbers verified live against the canonical API.
How Gemini 3.1 Pro compares within the Gemini family
Google currently ships six Gemini-family models in our live lineup. Gemini 3.1 Pro is the new mid-tier flagship — cheaper than the top-tier 3.5 Pro, smarter than the Flash tiers, and a meaningful upgrade over the year-old 2.5 Pro at the same list price.
| Model | Input | Output | Context | Max out | II | Cutoff | Blended $/M |
|---|---|---|---|---|---|---|---|
| Gemini 3.1 Pro | $1.25 | $10 | 1M | 65,536 | 57 | Feb 2026 | $3.88 |
| Gemini 3.5 Pro | $1.50 | $12 | 2M | 65,536 | 63 | Apr 2026 | $4.65 |
| Gemini 3.5 Flash | $0.30 | $2.50 | 1M | 65,536 | 55 | Feb 2026 | $0.96 |
| Gemini 2.5 Pro | $1.25 | $10 | 2M | 65,536 | 53 | Jan 2025 | $3.88 |
| Gemini 2.5 Flash | $0.30 | $2.50 | 1M | 65,536 | 48 | Jan 2025 | $0.96 |
Source: LLMCalculator live API. Blended $/M = 0.7 × input + 0.3 × output, per 1M total tokens.
How Gemini 3.1 Pro compares to the frontier
Across the current frontier peers, Gemini 3.1 Pro lands in the middle of the pack on raw price — tied on input with GPT-5.5, well under the Claude tier, and well above the cheapest "smart" tier (Grok 4.3, DeepSeek V4 Pro).
| Model | Input | Output | Blended $/M | vs Gemini 3.1 Pro |
|---|---|---|---|---|
| Gemini 3.1 Pro | $1.25 | $10 | $3.88 | — |
| GPT-5.5 (xhigh) | $1.25 | $10 | $3.88 | Same list price |
| Claude Sonnet 5 | $3 | $15 | $6.60 | +70% (Gemini 3.1 Pro ~41% cheaper) |
| Claude Opus 4.8 | $5 | $25 | $11.00 | +184% (Gemini 3.1 Pro ~65% cheaper) |
| Grok 4.3 (high) | $0.30 | $1.50 | $0.66 | −83% (cheaper, lower II) |
| DeepSeek V4 Pro | $0.27 | $1.10 | $0.52 | −87% (cheaper, lower II) |
Source: LLMCalculator live API. Blended $/M = 0.7 × input + 0.3 × output.
Workload cost examples
Three realistic workload shapes, per-batch cost. Gemini 3.1 Pro tracks GPT-5.5 exactly because they share list price; the only model in this comparison cheaper is the lower-intelligence Gemini 3.5 Flash.
| Workload | In / Out tokens | Gemini 3.1 Pro | GPT-5.5 (xhigh) | Sonnet 5 | 3.5 Flash |
|---|---|---|---|---|---|
| Short chatbot | 200 / 50 | $0.75 / 1k reqs | $0.75 / 1k reqs | $1.35 / 1k reqs | $0.18 / 1k reqs |
| RAG query | 5,000 / 500 | $11.25 / 1k reqs | $11.25 / 1k reqs | $22.50 / 1k reqs | $2.75 / 1k reqs |
| Long-doc analysis | 50,000 / 1,000 | $72.50 / 100 docs | $72.50 / 100 docs | $165.00 / 100 docs | $17.50 / 100 docs |
Source: LLMCalculator live API. Per-token list price; no caching, batching, or volume discounts applied. Run your own numbers on the Gemini pricing calculator.
When Gemini 3.1 Pro is the right pick
- Reasoning + long-context on a mid-tier budget. 1M context + 65,536 max output + 57 II at $1.25/$10 is the strongest pure-value reasoning slot in Google's current lineup.
- Vision workloads. Gemini 3.1 Pro is multimodal out of the box (the live lineup lists "reasoning, vision, long-context" as its best-for tags), so it handles image+text inputs without separate vision pricing.
- Migrating off Gemini 2.5 Pro. Same list price, +4 II points, knowledge cutoff 13 months newer (Feb 2026 vs Jan 2025). The only thing 2.5 Pro still has over 3.1 is a 2M context window vs 1M — if you actually need 2M context, stay on 2.5 Pro (or step up to 3.5 Pro).
When to pick something else
- Need the cheapest "smart enough" model? Gemini 3.5 Flash at $0.30/$2.50 is ~75% cheaper blended, only 2 II points behind, and shares Gemini 3.1 Pro's 1M context and 65,536 max output.
- Need the absolute cheapest frontier-tier capability? Grok 4.3 high ($0.30/$1.50) and DeepSeek V4 Pro ($0.27/$1.10) undercut Gemini 3.1 Pro by 83–87% on blended cost, with an II tradeoff.
- Need 2M+ context? Gemini 3.5 Pro (2M ctx, ii 63) and Gemini 2.5 Pro (2M ctx, ii 53) are your Gemini-side options at $1.50/$12 and $1.25/$10 respectively.
- Need top-of-frontier reasoning? Claude Fable 5 (ii 66) at $10/$50 is ~5.7× Gemini 3.1 Pro's blended cost. Claude Opus 4.8 ($5/$25, ii 62) is ~2.8×. Worth it only if you can measure the quality delta on your workload.
How to use Gemini 3.1 Pro cheaply in production
- Cache repeated context. If your RAG prompt reuses the same retrieved-context block across many calls, Google's implicit caching discounts kick in on hits — measured per-cache-hit pricing, not per-token list.
- Right-size your context window. 1M tokens is large; if your actual workload only needs 50k, you're not paying more — but a smaller-context model (e.g. Gemini 3.5 Flash at $0.30/$2.50) may produce equivalent quality at a fraction of the cost.
- Route by complexity. Use Gemini 3.1 Pro for the hard reasoning turn and Gemini 3.5 Flash for the easy follow-ups. A simple classifier + dual-model routing typically cuts blended spend 40–60%.
FAQ
How much does Gemini 3.1 Pro cost per million tokens?
Gemini 3.1 Pro costs $1.25 per 1M input tokens and $10 per 1M output tokens at list price. On a typical 70% input / 30% output conversation blend, that's $3.88 per 1M total tokens.
What is the context window of Gemini 3.1 Pro?
Gemini 3.1 Pro has a 1,000,000-token context window and a 65,536-token max output. If you need 2M context, step up to Gemini 3.5 Pro (ii 63) or stay on the older Gemini 2.5 Pro (ii 53, January 2025 cutoff).
Is Gemini 3.1 Pro cheaper than GPT-5.5?
At list price they're identical ($1.25/$10 per 1M tokens, $3.88 blended). GPT-5.5 has a slightly higher intelligence index (60 vs 57) and a 922k context vs 1M for Gemini 3.1 Pro; Gemini 3.1 Pro has a 65,536 max output vs 32,768 for GPT-5.5 and a Feb 2026 knowledge cutoff vs Oct 2025. Pick on workload fit, not price.
Should I use Gemini 3.1 Pro or Gemini 3.5 Pro?
Gemini 3.1 Pro is ~17% cheaper blended ($3.88 vs $4.65 per 1M tokens) and shares the 65,536 max output. Gemini 3.5 Pro has a 6-point higher intelligence index (63 vs 57), a 2M context window vs 1M, and an April 2026 knowledge cutoff vs Feb 2026. Pick 3.1 Pro when cost matters and 1M context is enough; pick 3.5 Pro when you need the bigger context or the absolute frontier of Google's lineup.
Sources & further reading
- Google Gemini pricing (official): ai.google.dev/pricing
- LLMCalculator live pricing API: /api/v1/models
- Gemini pricing calculator: /gemini-pricing-calculator