# Gemini 3.1 Pro Preview vs GPT-5.1

> Gemini 3.1 Pro Preview is the stronger model overall, scoring 56.7 to 48.8 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gemini-3-1-pro-preview-vs-gpt-5-1
- Last updated: 2026-10-10
- Shared benchmarks: 49

## Summary

- They share 49 benchmarks with published results for both. Gemini 3.1 Pro Preview scores higher in 6 categories and GPT-5.1 in 4 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Gemini 3.1 Pro Preview leads 71.7 to 39.9.
- The biggest single-benchmark swing is ARC-AGI-2: 77.1% for Gemini 3.1 Pro Preview and 17.6% for GPT-5.1.
- GPT-5.1 is cheaper at $1.25 / $10 per million input/output tokens, against $2 / $12 for Gemini 3.1 Pro Preview.
- Gemini 3.1 Pro Preview accepts more context: 1.05M tokens versus 400K.

## Snapshot

| | Gemini 3.1 Pro Preview | GPT-5.1 |
|---|---|---|
| Provider | Google | OpenAI |
| Noometry Index | 56.7 | 48.8 |
| Rank | 20 | 47 |
| Context | 1.05M | 400K |
| Input $/M | $2 | $1.25 |
| Output $/M | $12 | $10 |
| Weights | Proprietary | Proprietary |

## Coding

- Gemini 3.1 Pro Preview: 42.7 (#88)
- GPT-5.1: 46.9 (#59)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5.1 |
|---|---|---|
| SWE-bench Verified | 75.6% | 68% |
| LMArena WebDev | 1446 | 1395 |
| SciCode | 58.9% | 43.3% |
| GSO | 22.6% | 13.7% |
| WeirdML | 72.1% | 60.8% |
| LMArena Coding | 1483 | 1454 |
| ALE-Bench | 1,161 | 1,192 |
| DeepSWE | 11.7% | — |
| SWE-bench Verified (bash only) | — | 66% |
| LiveBench Coding | — | 72.5% |
| MirrorCode | 8.9% | — |
| AlgoTune | 2.02 | — |

## Agentic & Tool Use

- Gemini 3.1 Pro Preview: 37.7 (#33)
- GPT-5.1: 32.5 (#55)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5.1 |
|---|---|---|
| Terminal-Bench | 80.2% | 47.6% |
| DeepResearch Bench | 47.8% | 42.8% |
| LMArena Search | 1211 | 1199 |
| Vending-Bench 2 | 3,774 | 1,473 |
| APEX-Agents | 35.3% | — |
| τ²-bench Banking | 26% | — |
| PostTrainBench | 22% | — |
| BALROG | 57% | — |
| ExploitBench | 26.1% | — |
| GBAEval | 0.8% | — |
| GDP.pdf | 17% | — |
| METR Time Horizons | 77% | — |

## Reasoning

- Gemini 3.1 Pro Preview: 71.7 (#9)
- GPT-5.1: 39.9 (#50)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5.1 |
|---|---|---|
| ARC-AGI-2 | 77.1% | 17.6% |
| SimpleBench | 79.6% | 53.2% |
| ARC-AGI-1 | 98% | 72.8% |
| CritPt | 17.7% | 4.9% |
| Chess Puzzles | 55% | 32% |
| EnigmaEval | 36.8% | 11.2% |
| LMArena Hard Prompts | 1484 | 1457 |
| Mystery Game Puzzles | 34% | 19% |
| DTBench | 97.1% | 90.1% |
| LMCA | 53.8% | 43.9% |
| Epoch Capabilities Index | 154.92 | 149.66 |
| ForecastBench | 59 | 58.1 |
| NYT Connections (extended) | 97.4% | — |
| Thematic Generalization | 79.4% | — |
| EBR-Bench | 14.3% | — |
| LiveBench Reasoning | — | 95.8% |
| LiveBench Data Analysis | — | 72.1% |
| LiveBench | — | 78.8% |

## Math

- Gemini 3.1 Pro Preview: 61.8 (#28)
- GPT-5.1: 50.3 (#45)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5.1 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 95.6% | 88.6% |
| LMArena Math | 1485 | 1446 |
| FrontierMath (Feb 2025 set) | 36.9% | 31% |
| FrontierMath Tier 4 (v1) | 16.7% | 12.5% |
| FrontierMath (Tiers 1-3) | 59.6% | — |
| FrontierMath Tier 4 | 26.8% | — |
| MathArena Final-Answer Competitions | 86.5% | — |
| ProofBench | 26% | — |
| Omni-MATH | — | 46.4% |
| LiveBench Math | — | 94.5% |

## Knowledge

- Gemini 3.1 Pro Preview: 71.8 (#3)
- GPT-5.1: 50.5 (#65)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5.1 |
|---|---|---|
| GPQA Diamond | 94.4% | 87.6% |
| Humanity's Last Exam | 46.4% | 23.7% |
| SimpleQA Verified | 73.5% | 48% |
| Vectara Hallucination Rate | 10.4% | 10.9% |
| LMArena Expert | 1484 | 1470 |
| MMLU-Pro | — | 57.9% |
| GPQA (HELM) | — | 44.2% |

## Multimodal

- Gemini 3.1 Pro Preview: 38.1 (#60)
- GPT-5.1: 44.0 (#17)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5.1 |
|---|---|---|
| LMArena Vision | 1296 | 1250 |
| LMArena Document | 1444 | 1403 |
| VPCT | — | 58.7% |
| Blueprint-Bench 2 | 26.5% | — |
| Furniture Assembly | 26.7% | — |

## Multilingual

- Gemini 3.1 Pro Preview: 57.1 (#11)
- GPT-5.1: 54.0 (#51)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5.1 |
|---|---|---|
| LMArena Non-English | 1475 | 1431 |
| LMArena Chinese | 1528 | 1495 |
| LMArena French | 1488 | 1451 |
| LMArena German | 1491 | 1439 |
| LMArena Japanese | 1494 | 1453 |
| LMArena Korean | 1458 | 1401 |
| LMArena Russian | 1496 | 1435 |
| LMArena Spanish | 1480 | 1430 |

## Instruction Following

- Gemini 3.1 Pro Preview: 77.1 (#31)
- GPT-5.1: 84.0 (#1)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5.1 |
|---|---|---|
| LMArena Instruction Following | 1465 | 1442 |
| LiveBench Instruction Following | — | 93.3% |
| IFEval | — | 93.5% |

## Long Context

- Gemini 3.1 Pro Preview: 47.3 (#17)
- GPT-5.1: 47.6 (#14)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5.1 |
|---|---|---|
| CL-bench | 20.8% | 23.7% |
| CL-bench Life | 16.9% | 17.3% |
| LMArena Longer Query | 1482 | 1448 |

## Writing & Preference

- Gemini 3.1 Pro Preview: 66.3 (#36)
- GPT-5.1: 64.7 (#53)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5.1 |
|---|---|---|
| LMArena Text | 1480 | 1443 |
| LMArena Creative Writing | 1482 | 1427 |
| LMArena Multi-Turn | 1486 | 1449 |
| EQ-Bench Creative Writing | 1491 | — |
| WildBench | — | 86.3% |
| EQ-Bench 4 | 1142 | — |
| LiveBench Language | — | 80.2% |

## FAQ

### Is Gemini 3.1 Pro Preview better than GPT-5.1?

Gemini 3.1 Pro Preview is the stronger model overall, scoring 56.7 to 48.8 on the Noometry Index.

### Which is cheaper, Gemini 3.1 Pro Preview or GPT-5.1?

GPT-5.1 is cheaper. It lists at $1.25 per million input tokens and $10 per million output tokens; Gemini 3.1 Pro Preview lists at $2 and $12.

### Is Gemini 3.1 Pro Preview or GPT-5.1 better for coding?

GPT-5.1 scores higher on coding benchmarks: 46.9 versus 42.7 in the Noometry coding category.

### Which has the bigger context window?

Gemini 3.1 Pro Preview does, with 1.05M tokens against 400K.

### How many benchmarks do Gemini 3.1 Pro Preview and GPT-5.1 share?

49 benchmarks have published results for both models. Gemini 3.1 Pro Preview has 71 scored results on Noometry and GPT-5.1 has 63.