# Gemini 3.1 Pro Preview Usage, Cost & Rank | OpenCode Data

> Gemini 3.1 Pro Preview costs $2.00 per 1M input tokens and $12.00 per 1M output tokens.

Reasoning-first Gemini preview for agentic coding and complex problem solving

- Page: https://opencode.ai/data/google/gemini-3-1-pro-preview
- JSON: https://opencode.ai/data/google/gemini-3-1-pro-preview.json
- Model ID: google/gemini-3.1-pro-preview
- Lab: Google

## Model facts

| Fact | Value |
| --- | --- |
| Context window | 1M |
| Max output | 66K |
| Knowledge cutoff | 2025-01 |
| Release date | 2026-02-19 |
| Input modalities | text, image, video, audio, pdf |
| Output modalities | text |
| Reasoning | Yes |
| Tool calling | Yes |
| Open weights | No |

## Pricing: USD per 1M tokens

| Input | Output | Cached input | Cache write |
| --- | --- | --- | --- |
| $2.00 | $12.00 | $0.20 | - |

## Benchmarks

| Benchmark | Score | Metric | Source |
| --- | --- | --- | --- |
| SWE-Bench Pro | 54.2 | resolve rate | https://www.anthropic.com/news/claude-opus-4-8 |
| Terminal-Bench | 70.3 | success rate | https://www.anthropic.com/news/claude-opus-4-8 |
| SWE-Bench Pro | 46.1 | resolve rate | https://labs.scale.com/leaderboard/swe_bench_pro_public |
| SWE-Atlas Codebase QnA | 13.5 | score | https://labs.scale.com/leaderboard/sweatlas-qna |
| SWE-Atlas Refactoring | 33.81 | score | https://labs.scale.com/leaderboard/sweatlas-refactoring |
| SWE-Atlas Test Writing | 29.84 | score | https://labs.scale.com/leaderboard/sweatlas-tw |
| Artificial Analysis Coding Agent Index | 43 | average pass@1 | https://artificialanalysis.ai/agents/coding-agents |
| SWE-Atlas Codebase QnA | 45.6 | pass@1 | https://artificialanalysis.ai/agents/coding-agents |
| SWE-Bench Pro | 15.1 | pass@1 | https://artificialanalysis.ai/agents/coding-agents |
| Terminal-Bench | 68.3 | pass@1 | https://artificialanalysis.ai/agents/coding-agents |
| GPQA Diamond | 94.3 | accuracy | https://openai.com/index/introducing-gpt-5-5/ |
| Humanity's Last Exam | 44.4 | accuracy | https://deepmind.google/models/gemini/flash/ |
| ARC-AGI-2 | 77.1 | accuracy | https://deepmind.google/models/gemini/flash/ |
| MMMU Pro | 80.5 | accuracy | https://deepmind.google/models/gemini/flash/ |
| MCP Atlas | 78.2 | success rate | https://deepmind.google/models/gemini/flash/ |
| OSWorld-Verified | 76.2 | success rate | https://deepmind.google/models/gemini/flash/ |
| CharXiv Reasoning | 83.3 | accuracy | https://deepmind.google/models/gemini/flash/ |
| GDPval-AA | 1314 | Elo | https://deepmind.google/models/gemini/flash/ |

## Methodology

- Updates: Aggregated every hour. Days and weeks use UTC.
- Tokens: Input, output, reasoning, and cached tokens for each request.
- Users and sessions: Approximate counts of distinct users and OpenCode sessions.
- Cost: Session cost is the average cost per OpenCode session. Token prices are list prices from the OpenCode model catalog.
- Retention: The share of a model's users in one week who use it again the next week.
- Citation: Cite OpenCode Data (opencode.ai/data) with the update time shown at the top of the page.