# GPT-5.2 vs Mistral Small 3.1

> GPT-5.2 is the stronger model overall, scoring 54.1 to 31.4 on the Noometry Index. Mistral Small 3.1 costs 12× less per token, which makes it the better buy when GPT-5.2's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/gpt-5-2-vs-mistral-small-3-1
- Last updated: 2026-10-10
- Shared benchmarks: 23

## Summary

- They share 23 benchmarks with published results for both. GPT-5.2 scores higher in 9 categories and Mistral Small 3.1 in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.2 leads 59.8 to 13.7.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 96.1% for GPT-5.2 and 3.9% for Mistral Small 3.1.
- Mistral Small 3.1 is cheaper at $0.35 / $0.56 per million input/output tokens, against $1.75 / $14 for GPT-5.2.
- GPT-5.2 accepts more context: 400K tokens versus 128K.
- Mistral Small 3.1 has downloadable open weights; the other is API-only.

## Snapshot

| | GPT-5.2 | Mistral Small 3.1 |
|---|---|---|
| Provider | OpenAI | Mistral AI |
| Noometry Index | 54.1 | 31.4 |
| Rank | 31 | 263 |
| Context | 400K | 128K |
| Input $/M | $1.75 | $0.35 |
| Output $/M | $14 | $0.56 |
| Weights | Proprietary | Open |

## Coding

- GPT-5.2: 52.2 (#32)
- Mistral Small 3.1: 37.7 (#175)

| Benchmark | GPT-5.2 | Mistral Small 3.1 |
|---|---|---|
| LMArena Coding | 1447 | 1309 |
| SWE-bench Verified | 73.8% | — |
| SWE-bench Verified (bash only) | 72.8% | — |
| LMArena WebDev | 1415 | — |
| SWE-bench Multilingual | 66.7% | — |
| GSO | 27.4% | — |
| WeirdML | 72.2% | — |
| ALE-Bench | 1,294 | — |
| AlgoTune | 2.05 | — |

## Agentic & Tool Use

- GPT-5.2: 40.0 (#24)
- Mistral Small 3.1: —

| Benchmark | GPT-5.2 | Mistral Small 3.1 |
|---|---|---|
| Terminal-Bench | 64.9% | — |
| Berkeley Function Calling Leaderboard | 55.9% | — |
| GDPval | 49.7% | — |
| Remote Labor Index | 2.5% | — |
| τ²-bench Airline | 83% | — |
| τ²-bench Banking | 32.2% | — |
| τ²-bench Retail | 81.6% | — |
| τ²-bench Telecom | 89.7% | — |
| DeepResearch Bench | 41.1% | — |
| LMArena Search | 1207 | — |
| METR Time Horizons | 75.3% | — |
| Vending-Bench 2 | 3,591 | — |

## Reasoning

- GPT-5.2: 50.1 (#33)
- Mistral Small 3.1: 18.9 (#249)

| Benchmark | GPT-5.2 | Mistral Small 3.1 |
|---|---|---|
| Chess Puzzles | 49% | 1% |
| LMArena Hard Prompts | 1445 | 1278 |
| Epoch Capabilities Index | 153.45 | 127.48 |
| ARC-AGI-2 | 52.9% | — |
| SimpleBench | 45.8% | — |
| Kagi LLM Benchmark | 73.3% | — |
| NYT Connections (extended) | 83.6% | — |
| ARC-AGI-1 | 86.2% | — |
| EnigmaEval | 10.4% | — |
| EBR-Bench | 23% | — |
| Mystery Game Puzzles | 23% | — |
| DTBench | 90.9% | — |
| LMCA | 43.9% | — |
| ForecastBench | 60.1 | — |

## Math

- GPT-5.2: 59.8 (#32)
- Mistral Small 3.1: 13.7 (#295)

| Benchmark | GPT-5.2 | Mistral Small 3.1 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 96.1% | 3.9% |
| LMArena Math | 1439 | 1261 |
| FrontierMath (Tiers 1-3) | 67.4% | — |
| FrontierMath Tier 4 | 31.7% | — |
| MathArena Final-Answer Competitions | 72% | — |
| ProofBench | 15% | — |
| Omni-MATH | — | 24.8% |
| FrontierMath (Feb 2025 set) | 40.7% | — |
| FrontierMath Tier 4 (v1) | 18.8% | — |

## Knowledge

- GPT-5.2: 59.2 (#29)
- Mistral Small 3.1: 22.6 (#266)

| Benchmark | GPT-5.2 | Mistral Small 3.1 |
|---|---|---|
| GPQA Diamond | 91.4% | 41.9% |
| LMArena Expert | 1445 | 1258 |
| Humanity's Last Exam | 27.8% | — |
| SimpleQA Verified | 37.1% | — |
| MMLU-Pro | — | 61% |
| Vectara Hallucination Rate | 8.4% | — |
| GPQA (HELM) | — | 39.2% |

## Multimodal

- GPT-5.2: 50.9 (#7)
- Mistral Small 3.1: 32.6 (#93)

| Benchmark | GPT-5.2 | Mistral Small 3.1 |
|---|---|---|
| LMArena Vision | 1268 | 1136 |
| VPCT | 84% | — |
| Furniture Assembly | 38.3% | — |
| LMArena Document | 1405 | — |

## Multilingual

- GPT-5.2: 53.6 (#63)
- Mistral Small 3.1: 41.4 (#204)

| Benchmark | GPT-5.2 | Mistral Small 3.1 |
|---|---|---|
| LMArena Non-English | 1425 | 1255 |
| LMArena Chinese | 1460 | 1254 |
| LMArena French | 1456 | 1270 |
| LMArena German | 1448 | 1266 |
| LMArena Japanese | 1421 | 1210 |
| LMArena Korean | 1390 | 1206 |
| LMArena Russian | 1440 | 1263 |
| LMArena Spanish | 1433 | 1283 |

## Instruction Following

- GPT-5.2: 74.9 (#83)
- Mistral Small 3.1: 63.8 (#225)

| Benchmark | GPT-5.2 | Mistral Small 3.1 |
|---|---|---|
| LMArena Instruction Following | 1417 | 1264 |
| IFEval | — | 75% |

## Long Context

- GPT-5.2: 44.0 (#74)
- Mistral Small 3.1: 39.5 (#172)

| Benchmark | GPT-5.2 | Mistral Small 3.1 |
|---|---|---|
| LMArena Longer Query | 1428 | 1300 |
| CL-bench | 18.2% | — |

## Writing & Preference

- GPT-5.2: 67.0 (#30)
- Mistral Small 3.1: 37.2 (#254)

| Benchmark | GPT-5.2 | Mistral Small 3.1 |
|---|---|---|
| LMArena Text | 1439 | 1278 |
| LMArena Creative Writing | 1402 | 1253 |
| EQ-Bench Creative Writing | 1703 | 761 |
| LMArena Multi-Turn | 1458 | 1270 |
| WildBench | — | 78.8% |

## FAQ

### Is GPT-5.2 better than Mistral Small 3.1?

GPT-5.2 is the stronger model overall, scoring 54.1 to 31.4 on the Noometry Index. Mistral Small 3.1 costs 12× less per token, which makes it the better buy when GPT-5.2's lead doesn't matter for your workload.

### Which is cheaper, GPT-5.2 or Mistral Small 3.1?

Mistral Small 3.1 is cheaper. It lists at $0.35 per million input tokens and $0.56 per million output tokens; GPT-5.2 lists at $1.75 and $14.

### Is GPT-5.2 or Mistral Small 3.1 better for coding?

GPT-5.2 scores higher on coding benchmarks: 52.2 versus 37.7 in the Noometry coding category.

### Which has the bigger context window?

GPT-5.2 does, with 400K tokens against 128K.

### How many benchmarks do GPT-5.2 and Mistral Small 3.1 share?

23 benchmarks have published results for both models. GPT-5.2 has 67 scored results on Noometry and Mistral Small 3.1 has 28.