# Gemini 3.1 Pro

> Source: Benchize (https://benchize.com) — https://benchize.com/models/gemini-3-1-pro. When citing or reusing this content, credit "Benchize" and link to https://benchize.com/models/gemini-3-1-pro.

Google

Canonical URL: https://benchize.com/models/gemini-3-1-pro
Last updated: Thu, 22 Oct 2026 05:30:58 GMT

Reproduction and quotation must name Benchize, link to the canonical HTML URL, and retain individual source citations.

Google

## Benchmarks

| Benchmark | Score | ± | Rank | Cost | Source |  |
| --- | --- | --- | --- | --- | --- | --- |
| [Vending-Bench 2](https://benchize.com/benchmarks/vendingbench%3Av2.md) | — | — | — | $911.21 | [vendingbench](https://andonlabs.com/_app/immutable/chunks/Cjn0_qj9.js) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `final_balance` \| 911.2139999999997 \| USD \| one_simulated_year \| Higher is better \| \| `geometric_mean_balance` \| 2081.8941591252906 \| USD \| positive_final_balances \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration Gemini 3.1 Pro thinkingEnabled Not reported ### Reported statistics confidenceHalfWidth95 1279.0011974887311 positiveBalanceRuns 2 runCount 5 standardDeviation 1459.1498065757328 standardError 652.5516313718016 ### Source [Original source](https://andonlabs.com/_app/immutable/chunks/Cjn0_qj9.js) Source vendingbench Original model label Gemini 3.1 Pro Observed 2026-10-21T16:36:40.592Z Evidence SHA-256 `1a0b043147050298c405bcb743d73e425d532464202231cd18c27f836fa5f223` Source locator `/rows/42` |  |  |  |  |  |  |
| [PerceptionBench](https://benchize.com/benchmarks/perceptionbench.md) | 56.2% | — | 4 / 14 | — | [Moonshot AI / PerceptionBench · ba032c06e9b6ee3679171ff6ba643b7a0cfebe2e](https://github.com/MoonshotAI/PerceptionBench) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 56.2 \| percent \| overall \| Higher is better \| \| `visual_relation_accuracy` \| 58.8 \| percent \| atomic_capability \| Higher is better \| \| `counting_accuracy` \| 56.9 \| percent \| atomic_capability \| Higher is better \| \| `attribute_accuracy` \| 61.8 \| percent \| atomic_capability \| Higher is better \| \| `depth_3d_accuracy` \| 50.0 \| percent \| atomic_capability \| Higher is better \| \| `localization_accuracy` \| 52.7 \| percent \| atomic_capability \| Higher is better \| \| `comparison_accuracy` \| 61.7 \| percent \| atomic_capability \| Higher is better \| \| `fine_grained_recognition_accuracy` \| 54.8 \| percent \| atomic_capability \| Higher is better \| \| `contextual_integration_accuracy` \| 61.2 \| percent \| atomic_capability \| Higher is better \| \| `ocr_accuracy` \| 64.3 \| percent \| atomic_capability \| Higher is better \| \| `hallucination_accuracy` \| 40.6 \| percent \| atomic_capability \| Higher is better \| ### Configuration judgeModel gpt-oss-120b promptProtocol unified reasoningBudgetPolicy highest_available ### Reported statistics sampleSize 3000 ### Source Score publisher [Moonshot AI / PerceptionBench](https://github.com/MoonshotAI/PerceptionBench) Collected via [perceptionbench](https://api.github.com/repos/MoonshotAI/PerceptionBench/contents/README.md?ref=ba032c06e9b6ee3679171ff6ba643b7a0cfebe2e) Publisher reference Referenced by the collection source; the upstream result has not been independently captured. Benchmark [https://api.github.com/repos/MoonshotAI/PerceptionBench/contents/README.md?ref=ba032c06e9b6ee3679171ff6ba643b7a0cfebe2e](https://api.github.com/repos/MoonshotAI/PerceptionBench/contents/README.md?ref=ba032c06e9b6ee3679171ff6ba643b7a0cfebe2e) [Collected via](https://api.github.com/repos/MoonshotAI/PerceptionBench/contents/README.md?ref=ba032c06e9b6ee3679171ff6ba643b7a0cfebe2e) Source Moonshot AI / PerceptionBench Original model label Gemini-3.1-Pro Source version ba032c06e9b6ee3679171ff6ba643b7a0cfebe2e Observed 2026-10-21T16:33:36.858Z Evidence SHA-256 `0863d20bc397e2324ea76099fe703aed419dc7f0dbb359e522f7dac07dda6529` Source locator `/leaderboard/rows/3` |  |  |  |  |  |  |
| [PRBench Legal](https://benchize.com/benchmarks/scale%3Aprbench-legal.md) | 44.02source score | CI 42.9–45.14 | 17 / 26 | — | [scale](https://labs.scale.com/leaderboard/prbench-legal) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 44.0231 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration gemini-3.1-pro sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 42.9–45.14 contaminationWarning Not reported rank 24 reportedConfidenceHalfWidth 1.1208 ### Source [Original source](https://labs.scale.com/leaderboard/prbench-legal) Source scale Original model label gemini-3.1-pro Observed 2026-10-22T03:30:10.599Z Evidence SHA-256 `c086b601b5cff78b7934cfbcf79f15db48cd423c4e62e56a185b447b03c28f3c` Source locator `/rows/23` |  |  |  |  |  |  |
| [Long-Horizon Terminal-Bench](https://benchize.com/benchmarks/lhtb%3Aoverall.md) | 27.9% | — | 14 / 19 | $7.61 | [lhtb](https://zli12321.github.io/LHTB/script.js) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `mean_reward` \| 0.279 \| score_fraction \| reported \| Higher is better \| \| `solved_count` \| 2 \| tasks \| reward_gte_0_95 \| Higher is better \| \| `cost_per_task` \| 7.61 \| USD \| per_task \| Lower is better \| ### Configuration harness official_common_harness Reasoning effort unknown reportOrigin paper ### Reported statistics sampleSize 46 sourceThresholdCounts Not reported ### Source [Original source](https://zli12321.github.io/LHTB/script.js) Source lhtb Original model label Gemini 3.1 Pro Observed 2026-10-22T05:00:05.057Z Evidence SHA-256 `8b9c27664cc6298fa234c607ca27c953ec91b2eb9d116b8530cfaaa6722fa39e` Source locator `/rows/15` |  |  |  |  |  |  |
| [OfficeQA Pro V2 — Agent Harness](https://benchize.com/benchmarks/officeqa-pro%3Av2-agent.md) | 14.4% | — | 6 / 7 | — | [officeqa pro · sha256:d97762a15ad277575c2f567c4e3a318b6bb1d4e4d4d0008cd4f5c7953bae3253](https://api.github.com/repos/databricks/officeqa/git/blobs/cebf9bca23960fa1328f3421f0cb28d236cd86eb) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `correctness` \| 14.4 \| percent \| agent_full_tools \| Higher is better \| ### Configuration harness Gemini Antigravity ### Reported statistics sampleSize 90 ### Source [Original source](https://api.github.com/repos/databricks/officeqa/git/blobs/cebf9bca23960fa1328f3421f0cb28d236cd86eb) Source officeqa pro Original model label Gemini 3.1 Pro Source version sha256:d97762a15ad277575c2f567c4e3a318b6bb1d4e4d4d0008cd4f5c7953bae3253 Observed 2026-10-21T16:34:42.395Z Evidence SHA-256 `0e566748b11c0bfc76d6076b7d871f45fe43266e4787dfe523c359f3db0b7a10` Source locator `/chart-labels/0` |  |  |  |  |  |  |
| [HiL-Bench (Human-in-Loop Benchmark)](https://benchize.com/benchmarks/scale%3Ahil.md) | 35.33source score | CI 29.76–40.9 | 16 / 22 | — | [scale](https://labs.scale.com/leaderboard/hil) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 35.33 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration Gemini 3.1 Pro sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 29.76–40.9 contaminationWarning Not reported rank 22 reportedConfidenceHalfWidth 5.57 ### Source [Original source](https://labs.scale.com/leaderboard/hil) Source scale Original model label Gemini 3.1 Pro Observed 2026-10-22T03:30:07.512Z Evidence SHA-256 `fb954dc9c4a01e066526b773b8a269f2c27e73cfb0f1a943ec328ecd5be154c1` Source locator `/rows/22` |  |  |  |  |  |  |
| [DistressBench](https://benchize.com/benchmarks/scale%3Adistressbench.md) | 68.76source score | CI 67.46–70.06 | 19 / 27 | — | [scale](https://labs.scale.com/leaderboard/distressbench) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 68.76 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration gemini-3.1-pro sourceModelVersion gemini-3.1-pro-preview thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 67.46–70.06 contaminationWarning Not reported rank 24 reportedConfidenceHalfWidth 1.3 ### Source [Original source](https://labs.scale.com/leaderboard/distressbench) Source scale Original model label gemini-3.1-pro Observed 2026-10-22T03:30:05.896Z Evidence SHA-256 `6454425cf2678fa1119e99bffcdc827763d2b9c485395cb7aba2ef2e1f0cafbd` Source locator `/rows/23` |  |  |  |  |  |  |
| [PRBench Finance](https://benchize.com/benchmarks/scale%3Aprbench-finance.md) | 41.87source score | CI 40.66–43.09 | 16 / 22 | — | [scale](https://labs.scale.com/leaderboard/prbench-finance) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 41.8728 \| source_score \| reported \| Higher is better \| ### Configuration Reasoning effort unknown Model configuration gemini-3.1-pro sourceModelVersion Not reported thinkingEnabled Not reported ### Reported statistics Uncertainty Confidence interval: 40.66–43.09 contaminationWarning Not reported rank 25 reportedConfidenceHalfWidth 1.2163 ### Source [Original source](https://labs.scale.com/leaderboard/prbench-finance) Source scale Original model label gemini-3.1-pro Observed 2026-10-22T03:30:10.417Z Evidence SHA-256 `52572a2526d41cba3e52146e6063412a3850897b7608a0745f62baae712ab29b` Source locator `/rows/24` |  |  |  |  |  |  |
| [SpreadsheetBench v2-full](https://benchize.com/benchmarks/spreadsheetbench%3Av2-full.md) | 23.68% | — | 3 / 6 | — | [spreadsheetbench](https://spreadsheetbench.github.io/data/leaderboard-v2-full.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 23.68 \| percent \| reported \| Higher is better \| \| `Template` \| 28.87 \| percent \| task_category \| Higher is better \| \| `Fin.Model` \| 31 \| percent \| task_category \| Higher is better \| \| `Debug` \| 7 \| percent \| task_category \| Higher is better \| \| `Visualization` \| 41.67 \| percent \| task_category \| Higher is better \| ### Configuration Reasoning effort unknown scaffold SWE-agent verifiedSubmission true ### Source [Original source](https://spreadsheetbench.github.io/data/leaderboard-v2-full.json) Source spreadsheetbench Original model label Gemini 3.1 Pro Observed 2026-10-22T04:20:04.823Z Evidence SHA-256 `56490b18fd005f51e088710996636a4de82054809673f24cca40200915be338a` Source locator `/rows/5` |  |  |  |  |  |  |

---
Cite as: Gemini 3.1 Pro — Benchize, https://benchize.com/models/gemini-3-1-pro (retrieved 2026-10-22)