Model comparison
Mistral Medium vs Qwen2.5 72B Instruct
Mistral Medium is the stronger model overall, scoring 36.3 to 31.9 on the Noometry Index.
Last verified . 23 shared benchmarks.
Mistral AI
Alibaba (Qwen)Summary
- They share 23 benchmarks with published results for both. Mistral Medium scores higher in 8 categories and Qwen2.5 72B Instruct in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Mistral Medium leads 60.0 to 46.7.
- The biggest single-benchmark swing is WeirdML: 43.7% for Mistral Medium and 16% for Qwen2.5 72B Instruct.
- Qwen2.5 72B Instruct is cheaper at $1.40 / $5.60 per million input/output tokens, against $1.50 / $7.50 for Mistral Medium.
- Mistral Medium accepts more context: 262K tokens versus 131K.
Side by side
| Mistral Medium | Qwen2.5 72B Instruct | |
|---|---|---|
| Provider | Mistral AI | Alibaba (Qwen) |
| Noometry Index | 36.3 | 31.9 |
| Released | 2023-12-11 | 2024-09 |
| Weights | Open | Open |
| Context window | 262K | 131K |
| Max output | 262K | 8K |
| Input $ / M tokens | $1.50 | $1.40 |
| Output $ / M tokens | $7.50 | $5.60 |
| Results tracked | 36 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Mistral Medium: 34.2 (#243), Qwen2.5 72B Instruct: 33.2 (#260)
| Benchmark | Mistral Medium | Qwen2.5 72B Instruct |
|---|---|---|
| WeirdML | 43.7% | 16% |
| LMArena Coding | 1434 | 1292 |
| FrontierCode | 8% | — |
| SciCode | 40.2% | — |
| BigCodeBench Instruct | — | 45.8% |
| BigCodeBench Complete | — | 55.9% |
| ALE-Bench | 763.98 | — |
Agentic & Tool Use Mistral Medium leads
Mistral Medium: 28.3 (#90), Qwen2.5 72B Instruct: 22.1 (#133)
| Benchmark | Mistral Medium | Qwen2.5 72B Instruct |
|---|---|---|
| Berkeley Function Calling Leaderboard | 37.7% | — |
| TheAgentCompany | — | 5.7% |
| BALROG | — | 16.2% |
| METR Time Horizons | — | 35.8% |
Reasoning Mistral Medium leads
Mistral Medium: 24.0 (#167), Qwen2.5 72B Instruct: 22.3 (#199)
| Benchmark | Mistral Medium | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Hard Prompts | 1426 | 1271 |
| DTBench | 75.5% | 62.9% |
| LMCA | 26.1% | 13.4% |
| Kagi LLM Benchmark | 50% | — |
| CritPt | 0% | — |
| Surface Evolver Bench | 26.9% | — |
| BIG-Bench Hard | — | 79.8% |
| Epoch Capabilities Index | — | 129 |
| ForecastBench | — | 57.5 |
| HellaSwag | — | 84.8% |
| PIQA | — | 82.6% |
| WinoGrande | — | 82.3% |
Math Mistral Medium leads
Mistral Medium: 28.1 (#245), Qwen2.5 72B Instruct: 19.3 (#287)
| Benchmark | Mistral Medium | Qwen2.5 72B Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 32.2% | 8.1% |
| LMArena Math | 1408 | 1283 |
| MATH Level 5 | 81.6% | 63.2% |
| ProofBench | 9% | — |
| Omni-MATH | — | 33% |
| FrontierMath (Feb 2025 set) | 0.3% | — |
Knowledge Qwen2.5 72B Instruct leads
Mistral Medium: 25.0 (#265), Qwen2.5 72B Instruct: 27.0 (#253)
| Benchmark | Mistral Medium | Qwen2.5 72B Instruct |
|---|---|---|
| GPQA Diamond | 59.5% | 49.1% |
| LMArena Expert | 1408 | 1245 |
| Humanity's Last Exam | 4.5% | — |
| MMLU-Pro | — | 63.1% |
| Confabulations | — | 19.1% |
| Vectara Hallucination Rate | 22.7% | — |
| GPQA (HELM) | — | 42.6% |
| ARC (AI2) Challenge | — | 94.5% |
| MMLU | — | 85.3% |
| TriviaQA | — | 71.9% |
Multimodal Not comparable
Mistral Medium: 35.3 (#88), Qwen2.5 72B Instruct: —
| Benchmark | Mistral Medium | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Vision | 1172 | — |
Multilingual Mistral Medium leads
Mistral Medium: 52.1 (#91), Qwen2.5 72B Instruct: 41.0 (#213)
| Benchmark | Mistral Medium | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Non-English | 1408 | 1252 |
| LMArena Chinese | 1447 | 1272 |
| LMArena French | 1459 | 1280 |
| LMArena German | 1432 | 1234 |
| LMArena Japanese | 1378 | 1180 |
| LMArena Korean | 1380 | 1188 |
| LMArena Russian | 1411 | 1264 |
| LMArena Spanish | 1433 | 1256 |
Instruction Following Mistral Medium leads
Mistral Medium: 73.7 (#116), Qwen2.5 72B Instruct: 65.5 (#221)
| Benchmark | Mistral Medium | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Instruction Following | 1398 | 1254 |
| IFEval | — | 80.6% |
Long Context Mistral Medium leads
Mistral Medium: 42.9 (#114), Qwen2.5 72B Instruct: 38.9 (#188)
| Benchmark | Mistral Medium | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Longer Query | 1406 | 1282 |
Writing & Preference Mistral Medium leads
Mistral Medium: 60.0 (#103), Qwen2.5 72B Instruct: 46.7 (#215)
| Benchmark | Mistral Medium | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Text | 1424 | 1269 |
| LMArena Creative Writing | 1391 | 1221 |
| LMArena Multi-Turn | 1418 | 1272 |
| Short-Story Creative Writing | 77.3% | — |
| WildBench | — | 80.2% |
Frequently asked questions
Is Mistral Medium better than Qwen2.5 72B Instruct?
Mistral Medium is the stronger model overall, scoring 36.3 to 31.9 on the Noometry Index.
Which is cheaper, Mistral Medium or Qwen2.5 72B Instruct?
Qwen2.5 72B Instruct is cheaper. It lists at $1.40 per million input tokens and $5.60 per million output tokens; Mistral Medium lists at $1.50 and $7.50.
Is Mistral Medium or Qwen2.5 72B Instruct better for coding?
They score almost the same on coding (34.2 vs 33.2); test both on your own repository before choosing.
Which has the bigger context window?
Mistral Medium does, with 262K tokens against 131K.
How many benchmarks do Mistral Medium and Qwen2.5 72B Instruct share?
23 benchmarks have published results for both models. Mistral Medium has 36 scored results on Noometry and Qwen2.5 72B Instruct has 43.