# Gemma 4 31B IT vs GPT-5.1

> GPT-5.1 is the stronger model overall, scoring 48.8 to 43.2 on the Noometry Index. Gemma 4 31B IT costs 23× less per token, which makes it the better buy when GPT-5.1's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/gemma-4-31b-it-vs-gpt-5-1
- Last updated: 2026-10-10
- Shared benchmarks: 29

## Summary

- They share 29 benchmarks with published results for both. Gemma 4 31B IT scores higher in 0 categories and GPT-5.1 in 9 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5.1 leads 39.9 to 26.4.
- The biggest single-benchmark swing is SimpleQA Verified: 10.4% for Gemma 4 31B IT and 48% for GPT-5.1.
- Gemma 4 31B IT is cheaper at $0.09 / $0.34 per million input/output tokens, against $1.25 / $10 for GPT-5.1.
- GPT-5.1 accepts more context: 400K tokens versus 262K.
- Gemma 4 31B IT has downloadable open weights; the other is API-only.

## Snapshot

| | Gemma 4 31B IT | GPT-5.1 |
|---|---|---|
| Provider | Google | OpenAI |
| Noometry Index | 43.2 | 48.8 |
| Rank | 85 | 47 |
| Context | 262K | 400K |
| Input $/M | $0.09 | $1.25 |
| Output $/M | $0.34 | $10 |
| Weights | Open | Proprietary |

## Coding

- Gemma 4 31B IT: 41.9 (#98)
- GPT-5.1: 46.9 (#59)

| Benchmark | Gemma 4 31B IT | GPT-5.1 |
|---|---|---|
| LMArena WebDev | 1364 | 1395 |
| SciCode | 43.4% | 43.3% |
| WeirdML | 52.3% | 60.8% |
| LMArena Coding | 1459 | 1454 |
| ALE-Bench | 925.5 | 1,192 |
| SWE-bench Verified | — | 68% |
| SWE-bench Verified (bash only) | — | 66% |
| GSO | — | 13.7% |
| LiveBench Coding | — | 72.5% |

## Agentic & Tool Use

- Gemma 4 31B IT: —
- GPT-5.1: 32.5 (#55)

| Benchmark | Gemma 4 31B IT | GPT-5.1 |
|---|---|---|
| Terminal-Bench | — | 47.6% |
| DeepResearch Bench | — | 42.8% |
| LMArena Search | — | 1199 |
| Vending-Bench 2 | — | 1,473 |

## Reasoning

- Gemma 4 31B IT: 26.4 (#119)
- GPT-5.1: 39.9 (#50)

| Benchmark | Gemma 4 31B IT | GPT-5.1 |
|---|---|---|
| CritPt | 1.4% | 4.9% |
| Chess Puzzles | 5% | 32% |
| LMArena Hard Prompts | 1448 | 1457 |
| DTBench | 82.7% | 90.1% |
| LMCA | 39.3% | 43.9% |
| Epoch Capabilities Index | 142.74 | 149.66 |
| ARC-AGI-2 | — | 17.6% |
| SimpleBench | — | 53.2% |
| Kagi LLM Benchmark | 63.5% | — |
| NYT Connections (extended) | 70.6% | — |
| ARC-AGI-1 | — | 72.8% |
| EnigmaEval | — | 11.2% |
| Thematic Generalization | 53% | — |
| LiveBench Reasoning | — | 95.8% |
| Mystery Game Puzzles | — | 19% |
| LiveBench Data Analysis | — | 72.1% |
| Surface Evolver Bench | 30.6% | — |
| ForecastBench | — | 58.1 |
| LiveBench | — | 78.8% |

## Math

- Gemma 4 31B IT: 41.3 (#72)
- GPT-5.1: 50.3 (#45)

| Benchmark | Gemma 4 31B IT | GPT-5.1 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 73.3% | 88.6% |
| LMArena Math | 1464 | 1446 |
| Omni-MATH | — | 46.4% |
| LiveBench Math | — | 94.5% |
| FrontierMath (Feb 2025 set) | — | 31% |
| FrontierMath Tier 4 (v1) | — | 12.5% |

## Knowledge

- Gemma 4 31B IT: 37.9 (#141)
- GPT-5.1: 50.5 (#65)

| Benchmark | Gemma 4 31B IT | GPT-5.1 |
|---|---|---|
| GPQA Diamond | 75.8% | 87.6% |
| SimpleQA Verified | 10.4% | 48% |
| Vectara Hallucination Rate | 7.4% | 10.9% |
| LMArena Expert | 1465 | 1470 |
| Humanity's Last Exam | — | 23.7% |
| MMLU-Pro | — | 57.9% |
| GPQA (HELM) | — | 44.2% |

## Multimodal

- Gemma 4 31B IT: 41.0 (#32)
- GPT-5.1: 44.0 (#17)

| Benchmark | Gemma 4 31B IT | GPT-5.1 |
|---|---|---|
| LMArena Vision | 1278 | 1250 |
| LMArena Document | 1425 | 1403 |
| VPCT | — | 58.7% |

## Multilingual

- Gemma 4 31B IT: 54.0 (#52)
- GPT-5.1: 54.0 (#51)

| Benchmark | Gemma 4 31B IT | GPT-5.1 |
|---|---|---|
| LMArena Non-English | 1431 | 1431 |
| LMArena Chinese | 1478 | 1495 |
| LMArena French | 1436 | 1451 |
| LMArena Russian | 1462 | 1435 |
| LMArena Spanish | 1443 | 1430 |
| LMArena German | — | 1439 |
| LMArena Japanese | — | 1453 |
| LMArena Korean | — | 1401 |

## Instruction Following

- Gemma 4 31B IT: 75.6 (#58)
- GPT-5.1: 84.0 (#1)

| Benchmark | Gemma 4 31B IT | GPT-5.1 |
|---|---|---|
| LMArena Instruction Following | 1433 | 1442 |
| LiveBench Instruction Following | — | 93.3% |
| IFEval | — | 93.5% |

## Long Context

- Gemma 4 31B IT: 44.2 (#67)
- GPT-5.1: 47.6 (#14)

| Benchmark | Gemma 4 31B IT | GPT-5.1 |
|---|---|---|
| LMArena Longer Query | 1446 | 1448 |
| CL-bench | — | 23.7% |
| CL-bench Life | — | 17.3% |

## Writing & Preference

- Gemma 4 31B IT: 60.7 (#92)
- GPT-5.1: 64.7 (#53)

| Benchmark | Gemma 4 31B IT | GPT-5.1 |
|---|---|---|
| LMArena Text | 1443 | 1443 |
| LMArena Creative Writing | 1416 | 1427 |
| LMArena Multi-Turn | 1452 | 1449 |
| EQ-Bench Creative Writing | 1368 | — |
| WildBench | — | 86.3% |
| EQ-Bench 4 | 1120 | — |
| LiveBench Language | — | 80.2% |

## FAQ

### Is Gemma 4 31B IT better than GPT-5.1?

GPT-5.1 is the stronger model overall, scoring 48.8 to 43.2 on the Noometry Index. Gemma 4 31B IT costs 23× less per token, which makes it the better buy when GPT-5.1's lead doesn't matter for your workload.

### Which is cheaper, Gemma 4 31B IT or GPT-5.1?

Gemma 4 31B IT is cheaper. It lists at $0.09 per million input tokens and $0.34 per million output tokens; GPT-5.1 lists at $1.25 and $10.

### Is Gemma 4 31B IT or GPT-5.1 better for coding?

GPT-5.1 scores higher on coding benchmarks: 46.9 versus 41.9 in the Noometry coding category.

### Which has the bigger context window?

GPT-5.1 does, with 400K tokens against 262K.

### How many benchmarks do Gemma 4 31B IT and GPT-5.1 share?

29 benchmarks have published results for both models. Gemma 4 31B IT has 35 scored results on Noometry and GPT-5.1 has 63.