# Llama-3.3-70B-Instruct vs Mistral Medium

> Mistral Medium is the stronger model overall, scoring 36.3 to 30.6 on the Noometry Index. Llama-3.3-70B-Instruct costs 19× less per token, which makes it the better buy when Mistral Medium's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/llama-3-3-70b-instruct-vs-mistral-medium
- Last updated: 2026-10-12
- Shared benchmarks: 27

## Summary

- They share 27 benchmarks with published results for both. Llama-3.3-70B-Instruct scores higher in 1 category and Mistral Medium in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in long context, where Mistral Medium leads 42.9 to 26.4.
- The biggest single-benchmark swing is MATH Level 5: 41.6% for Llama-3.3-70B-Instruct and 81.6% for Mistral Medium.
- Llama-3.3-70B-Instruct is cheaper at $0.10 / $0.32 per million input/output tokens, against $1.50 / $7.50 for Mistral Medium.
- Mistral Medium accepts more context: 262K tokens versus 128K.

## Snapshot

| | Llama-3.3-70B-Instruct | Mistral Medium |
|---|---|---|
| Provider | Meta | Mistral AI |
| Noometry Index | 30.6 | 36.3 |
| Rank | 291 | 218 |
| Context | 128K | 262K |
| Input $/M | $0.10 | $1.50 |
| Output $/M | $0.32 | $7.50 |
| Weights | Open | Open |

## Coding

- Llama-3.3-70B-Instruct: 31.0 (#290)
- Mistral Medium: 34.2 (#243)

| Benchmark | Llama-3.3-70B-Instruct | Mistral Medium |
|---|---|---|
| SciCode | 26% | 40.2% |
| WeirdML | 14.4% | 43.7% |
| LMArena Coding | 1268 | 1434 |
| FrontierCode | — | 8% |
| BigCodeBench Instruct | 46.9% | — |
| LiveBench Coding | 36.6% | — |
| BigCodeBench Complete | 57.5% | — |
| ALE-Bench | — | 763.98 |

## Agentic & Tool Use

- Llama-3.3-70B-Instruct: 25.8 (#105)
- Mistral Medium: 28.3 (#90)

| Benchmark | Llama-3.3-70B-Instruct | Mistral Medium |
|---|---|---|
| Berkeley Function Calling Leaderboard | 31.9% | 37.7% |
| BALROG | 23% | — |

## Reasoning

- Llama-3.3-70B-Instruct: 14.1 (#327)
- Mistral Medium: 24.0 (#167)

| Benchmark | Llama-3.3-70B-Instruct | Mistral Medium |
|---|---|---|
| CritPt | 0% | 0% |
| LMArena Hard Prompts | 1257 | 1426 |
| DTBench | 59.5% | 75.5% |
| LMCA | 17.5% | 26.1% |
| SimpleBench | 19.9% | — |
| Kagi LLM Benchmark | — | 50% |
| LiveBench Reasoning | 50.8% | — |
| LiveBench Data Analysis | 49.5% | — |
| Surface Evolver Bench | — | 26.9% |
| Epoch Capabilities Index | 127.33 | — |
| ForecastBench | 58.6 | — |
| LiveBench | 50.2% | — |

## Math

- Llama-3.3-70B-Instruct: 15.3 (#298)
- Mistral Medium: 28.1 (#245)

| Benchmark | Llama-3.3-70B-Instruct | Mistral Medium |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 5.1% | 32.2% |
| LMArena Math | 1267 | 1408 |
| MATH Level 5 | 41.6% | 81.6% |
| ProofBench | — | 9% |
| LiveBench Math | 42.2% | — |
| FrontierMath (Feb 2025 set) | — | 0.3% |

## Knowledge

- Llama-3.3-70B-Instruct: 30.6 (#226)
- Mistral Medium: 25.0 (#265)

| Benchmark | Llama-3.3-70B-Instruct | Mistral Medium |
|---|---|---|
| GPQA Diamond | 47.4% | 59.5% |
| Vectara Hallucination Rate | 4.1% | 22.7% |
| LMArena Expert | 1225 | 1408 |
| Humanity's Last Exam | — | 4.5% |
| Confabulations | 22.8% | — |
| MMLU | 86.3% | — |

## Multimodal

- Llama-3.3-70B-Instruct: —
- Mistral Medium: 35.3 (#88)

| Benchmark | Llama-3.3-70B-Instruct | Mistral Medium |
|---|---|---|
| LMArena Vision | — | 1172 |

## Multilingual

- Llama-3.3-70B-Instruct: 39.9 (#220)
- Mistral Medium: 52.1 (#91)

| Benchmark | Llama-3.3-70B-Instruct | Mistral Medium |
|---|---|---|
| LMArena Non-English | 1236 | 1408 |
| LMArena Chinese | 1217 | 1447 |
| LMArena French | 1281 | 1459 |
| LMArena German | 1251 | 1432 |
| LMArena Japanese | 1150 | 1378 |
| LMArena Korean | 1143 | 1380 |
| LMArena Russian | 1252 | 1411 |
| LMArena Spanish | 1270 | 1433 |

## Instruction Following

- Llama-3.3-70B-Instruct: 71.1 (#157)
- Mistral Medium: 73.7 (#116)

| Benchmark | Llama-3.3-70B-Instruct | Mistral Medium |
|---|---|---|
| LMArena Instruction Following | 1242 | 1398 |
| LiveBench Instruction Following | 82.7% | — |

## Long Context

- Llama-3.3-70B-Instruct: 26.4 (#295)
- Mistral Medium: 42.9 (#114)

| Benchmark | Llama-3.3-70B-Instruct | Mistral Medium |
|---|---|---|
| LMArena Longer Query | 1256 | 1406 |
| Fiction.LiveBench | 33.3% | — |

## Writing & Preference

- Llama-3.3-70B-Instruct: 47.6 (#207)
- Mistral Medium: 60.0 (#103)

| Benchmark | Llama-3.3-70B-Instruct | Mistral Medium |
|---|---|---|
| LMArena Text | 1274 | 1424 |
| LMArena Creative Writing | 1250 | 1391 |
| LMArena Multi-Turn | 1280 | 1418 |
| Short-Story Creative Writing | — | 77.3% |
| LiveBench Language | 39.2% | — |

## FAQ

### Is Llama-3.3-70B-Instruct better than Mistral Medium?

Mistral Medium is the stronger model overall, scoring 36.3 to 30.6 on the Noometry Index. Llama-3.3-70B-Instruct costs 19× less per token, which makes it the better buy when Mistral Medium's lead doesn't matter for your workload.

### Which is cheaper, Llama-3.3-70B-Instruct or Mistral Medium?

Llama-3.3-70B-Instruct is cheaper. It lists at $0.10 per million input tokens and $0.32 per million output tokens; Mistral Medium lists at $1.50 and $7.50.

### Is Llama-3.3-70B-Instruct or Mistral Medium better for coding?

Mistral Medium scores higher on coding benchmarks: 34.2 versus 31.0 in the Noometry coding category.

### Which has the bigger context window?

Mistral Medium does, with 262K tokens against 128K.

### How many benchmarks do Llama-3.3-70B-Instruct and Mistral Medium share?

27 benchmarks have published results for both models. Llama-3.3-70B-Instruct has 43 scored results on Noometry and Mistral Medium has 36.