# Gemini 3.1 Pro Preview vs GPT-5 Mini

> Gemini 3.1 Pro Preview is the stronger model overall, scoring 56.7 to 41.7 on the Noometry Index. GPT-5 Mini costs 6.5× less per token, which makes it the better buy when Gemini 3.1 Pro Preview's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/gemini-3-1-pro-preview-vs-gpt-5-mini
- Last updated: 2026-10-10
- Shared benchmarks: 46

## Summary

- They share 46 benchmarks with published results for both. Gemini 3.1 Pro Preview scores higher in 10 categories and GPT-5 Mini in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Gemini 3.1 Pro Preview leads 71.7 to 24.1.
- The biggest single-benchmark swing is ARC-AGI-2: 77.1% for Gemini 3.1 Pro Preview and 4.4% for GPT-5 Mini.
- GPT-5 Mini is cheaper at $0.25 / $2 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 Mini |
|---|---|---|
| Provider | Google | OpenAI |
| Noometry Index | 56.7 | 41.7 |
| Rank | 20 | 115 |
| Context | 1.05M | 400K |
| Input $/M | $2 | $0.25 |
| Output $/M | $12 | $2 |
| Weights | Proprietary | Proprietary |

## Coding

- Gemini 3.1 Pro Preview: 42.7 (#88)
- GPT-5 Mini: 40.0 (#134)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5 Mini |
|---|---|---|
| SWE-bench Verified | 75.6% | 64.7% |
| SciCode | 58.9% | 39.2% |
| WeirdML | 72.1% | 52.7% |
| LMArena Coding | 1483 | 1406 |
| ALE-Bench | 1,161 | 799.77 |
| AlgoTune | 2.02 | 1.38 |
| DeepSWE | 11.7% | — |
| SWE-bench Verified (bash only) | — | 59.8% |
| LMArena WebDev | 1446 | — |
| SWE-bench Multilingual | — | 39.7% |
| GSO | 22.6% | — |
| MirrorCode | 8.9% | — |

## Agentic & Tool Use

- Gemini 3.1 Pro Preview: 37.7 (#33)
- GPT-5 Mini: 31.0 (#64)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5 Mini |
|---|---|---|
| Terminal-Bench | 80.2% | 34.8% |
| Vending-Bench 2 | 3,774 | -31.18 |
| APEX-Agents | 35.3% | — |
| Berkeley Function Calling Leaderboard | — | 55.5% |
| τ²-bench Banking | 26% | — |
| DeepResearch Bench | 47.8% | — |
| PostTrainBench | 22% | — |
| BALROG | 57% | — |
| ExploitBench | 26.1% | — |
| GBAEval | 0.8% | — |
| GDP.pdf | 17% | — |
| LMArena Search | 1211 | — |
| METR Time Horizons | 77% | — |

## Reasoning

- Gemini 3.1 Pro Preview: 71.7 (#9)
- GPT-5 Mini: 24.1 (#153)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5 Mini |
|---|---|---|
| ARC-AGI-2 | 77.1% | 4.4% |
| ARC-AGI-1 | 98% | 54.3% |
| CritPt | 17.7% | 0% |
| Chess Puzzles | 55% | 30% |
| EnigmaEval | 36.8% | 8.2% |
| LMArena Hard Prompts | 1484 | 1380 |
| Mystery Game Puzzles | 34% | 10% |
| DTBench | 97.1% | 80.5% |
| LMCA | 53.8% | 34.2% |
| Epoch Capabilities Index | 154.92 | 145.51 |
| ForecastBench | 59 | 61 |
| SimpleBench | 79.6% | — |
| Kagi LLM Benchmark | — | 70.3% |
| NYT Connections (extended) | 97.4% | — |
| Thematic Generalization | 79.4% | — |
| EBR-Bench | 14.3% | — |

## Math

- Gemini 3.1 Pro Preview: 61.8 (#28)
- GPT-5 Mini: 46.3 (#59)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5 Mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 59.6% | 46.7% |
| FrontierMath Tier 4 | 26.8% | 12.2% |
| OTIS Mock AIME 2024-2025 | 95.6% | 86.7% |
| ProofBench | 26% | 9% |
| LMArena Math | 1485 | 1377 |
| FrontierMath (Feb 2025 set) | 36.9% | 27.2% |
| FrontierMath Tier 4 (v1) | 16.7% | 6.3% |
| MathArena Final-Answer Competitions | 86.5% | — |
| Omni-MATH | — | 72.2% |
| MATH Level 5 | — | 97.8% |

## Knowledge

- Gemini 3.1 Pro Preview: 71.8 (#3)
- GPT-5 Mini: 45.5 (#80)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5 Mini |
|---|---|---|
| GPQA Diamond | 94.4% | 75% |
| Humanity's Last Exam | 46.4% | 19.4% |
| SimpleQA Verified | 73.5% | 21.6% |
| Vectara Hallucination Rate | 10.4% | 12.9% |
| LMArena Expert | 1484 | 1377 |
| MMLU-Pro | — | 83.5% |
| Confabulations | — | 13.3% |
| GPQA (HELM) | — | 75.6% |

## Multimodal

- Gemini 3.1 Pro Preview: 38.1 (#60)
- GPT-5 Mini: 35.0 (#80)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5 Mini |
|---|---|---|
| LMArena Vision | 1296 | 1202 |
| VPCT | — | 40.2% |
| Blueprint-Bench 2 | 26.5% | — |
| Furniture Assembly | 26.7% | — |
| LMArena Document | 1444 | — |

## Multilingual

- Gemini 3.1 Pro Preview: 57.1 (#11)
- GPT-5 Mini: 49.1 (#131)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5 Mini |
|---|---|---|
| LMArena Non-English | 1475 | 1363 |
| LMArena Chinese | 1528 | 1387 |
| LMArena French | 1488 | 1385 |
| LMArena German | 1491 | 1366 |
| LMArena Japanese | 1494 | 1342 |
| LMArena Korean | 1458 | 1306 |
| LMArena Russian | 1496 | 1362 |
| LMArena Spanish | 1480 | 1355 |

## Instruction Following

- Gemini 3.1 Pro Preview: 77.1 (#31)
- GPT-5 Mini: 76.4 (#43)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5 Mini |
|---|---|---|
| LMArena Instruction Following | 1465 | 1357 |
| IFEval | — | 92.7% |

## Long Context

- Gemini 3.1 Pro Preview: 47.3 (#17)
- GPT-5 Mini: 41.9 (#127)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5 Mini |
|---|---|---|
| LMArena Longer Query | 1482 | 1355 |
| Fiction.LiveBench | — | 69.4% |
| CL-bench | 20.8% | — |
| CL-bench Life | 16.9% | — |

## Writing & Preference

- Gemini 3.1 Pro Preview: 66.3 (#36)
- GPT-5 Mini: 55.4 (#143)

| Benchmark | Gemini 3.1 Pro Preview | GPT-5 Mini |
|---|---|---|
| LMArena Text | 1480 | 1373 |
| LMArena Creative Writing | 1482 | 1325 |
| EQ-Bench Creative Writing | 1491 | 1313 |
| LMArena Multi-Turn | 1486 | 1363 |
| Short-Story Creative Writing | — | 83.1% |
| WildBench | — | 85.5% |
| EQ-Bench 4 | 1142 | — |

## FAQ

### Is Gemini 3.1 Pro Preview better than GPT-5 Mini?

Gemini 3.1 Pro Preview is the stronger model overall, scoring 56.7 to 41.7 on the Noometry Index. GPT-5 Mini costs 6.5× less per token, which makes it the better buy when Gemini 3.1 Pro Preview's lead doesn't matter for your workload.

### Which is cheaper, Gemini 3.1 Pro Preview or GPT-5 Mini?

GPT-5 Mini is cheaper. It lists at $0.25 per million input tokens and $2 per million output tokens; Gemini 3.1 Pro Preview lists at $2 and $12.

### Is Gemini 3.1 Pro Preview or GPT-5 Mini better for coding?

Gemini 3.1 Pro Preview scores higher on coding benchmarks: 42.7 versus 40.0 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 Mini share?

46 benchmarks have published results for both models. Gemini 3.1 Pro Preview has 71 scored results on Noometry and GPT-5 Mini has 60.