Model comparison
Mistral 7B vs Qwen2.5 72B Instruct
Qwen2.5 72B Instruct is the stronger model overall, scoring 31.5 to 22.9 on the Noometry Index. Mistral 7B costs 9.8× less per token, which makes it the better buy when Qwen2.5 72B Instruct's lead doesn't matter for your workload.
Last verified . 30 shared benchmarks.
Mistral AI
Alibaba (Qwen)Summary
- They share 30 benchmarks with published results for both. Mistral 7B scores higher in 0 categories and Qwen2.5 72B Instruct in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Qwen2.5 72B Instruct leads 26.9 to 7.3.
- The biggest single-benchmark swing is MATH Level 5: 3.7% for Mistral 7B and 63.2% for Qwen2.5 72B Instruct.
- Mistral 7B is cheaper at $0.25 / $0.25 per million input/output tokens, against $1.40 / $5.60 for Qwen2.5 72B Instruct.
- Qwen2.5 72B Instruct accepts more context: 131K tokens versus 8K.
Side by side
| Mistral 7B | Qwen2.5 72B Instruct | |
|---|---|---|
| Provider | Mistral AI | Alibaba (Qwen) |
| Noometry Index | 22.9 | 31.5 |
| Released | 2023-09-27 | 2024-09 |
| Weights | Open | Open |
| Context window | 8K | 131K |
| Max output | 8K | 8K |
| Input $ / M tokens | $0.25 | $1.40 |
| Output $ / M tokens | $0.25 | $5.60 |
| Results tracked | 37 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen2.5 72B Instruct leads
Mistral 7B: 25.9 (#321), Qwen2.5 72B Instruct: 32.8 (#256)
| Benchmark | Mistral 7B | Qwen2.5 72B Instruct |
|---|---|---|
| BigCodeBench Instruct | 19.5% | 45.8% |
| LMArena Coding | 1082 | 1293 |
| BigCodeBench Complete | 27.3% | 55.9% |
| WeirdML | — | 16% |
| HumanEval+ | 36% | — |
| MBPP+ | 42.1% | — |
Agentic & Tool Use Not comparable
Mistral 7B: —, Qwen2.5 72B Instruct: 22.0 (#124)
| Benchmark | Mistral 7B | Qwen2.5 72B Instruct |
|---|---|---|
| TheAgentCompany | — | 5.7% |
| BALROG | — | 16.2% |
| METR Time Horizons | — | 35.8% |
Reasoning Qwen2.5 72B Instruct leads
Mistral 7B: 12.6 (#331), Qwen2.5 72B Instruct: 21.4 (#197)
| Benchmark | Mistral 7B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Hard Prompts | 1067 | 1271 |
| DTBench | 42.5% | 62.9% |
| BIG-Bench Hard | 56.1% | 79.8% |
| Epoch Capabilities Index | 111.97 | 129.01 |
| HellaSwag | 81% | 84.8% |
| PIQA | 83% | 82.6% |
| WinoGrande | 75.3% | 82.3% |
| Chess Puzzles | 0% | — |
| LMCA | — | 13.4% |
| Adversarial NLI | 47.1% | — |
| ForecastBench | — | 57.5 |
Math Qwen2.5 72B Instruct leads
Mistral 7B: 7.6 (#319), Qwen2.5 72B Instruct: 18.1 (#281)
| Benchmark | Mistral 7B | Qwen2.5 72B Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 0.3% | 8.1% |
| LMArena Math | 1085 | 1283 |
| MATH Level 5 | 3.7% | 63.2% |
| Omni-MATH | — | 33% |
| GSM8K | 54.4% | — |
Knowledge Qwen2.5 72B Instruct leads
Mistral 7B: 7.3 (#306), Qwen2.5 72B Instruct: 26.9 (#247)
| Benchmark | Mistral 7B | Qwen2.5 72B Instruct |
|---|---|---|
| GPQA Diamond | 15.2% | 49.1% |
| LMArena Expert | 1036 | 1245 |
| ARC (AI2) Challenge | 78.6% | 94.5% |
| MMLU | 62.5% | 85.3% |
| TriviaQA | 75.2% | 71.9% |
| MMLU-Pro | — | 63.1% |
| Confabulations | — | 19.1% |
| GPQA (HELM) | — | 42.6% |
| BoolQ | 87.4% | — |
| OpenBookQA | 79.8% | — |
Multilingual Qwen2.5 72B Instruct leads
Mistral 7B: 26.0 (#278), Qwen2.5 72B Instruct: 41.2 (#208)
| Benchmark | Mistral 7B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Non-English | 1012 | 1252 |
| LMArena Chinese | 1009 | 1272 |
| LMArena French | 1036 | 1279 |
| LMArena German | 987 | 1234 |
| LMArena Japanese | 878 | 1180 |
| LMArena Russian | 1018 | 1264 |
| LMArena Spanish | 1026 | 1255 |
| LMArena Korean | — | 1188 |
Instruction Following Qwen2.5 72B Instruct leads
Mistral 7B: 54.5 (#275), Qwen2.5 72B Instruct: 65.7 (#216)
| Benchmark | Mistral 7B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Instruction Following | 1060 | 1254 |
| IFEval | — | 80.6% |
Long Context Qwen2.5 72B Instruct leads
Mistral 7B: 32.2 (#266), Qwen2.5 72B Instruct: 38.9 (#183)
| Benchmark | Mistral 7B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Longer Query | 1060 | 1282 |
Writing & Preference Qwen2.5 72B Instruct leads
Mistral 7B: 30.9 (#281), Qwen2.5 72B Instruct: 46.9 (#210)
| Benchmark | Mistral 7B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Text | 1090 | 1269 |
| LMArena Creative Writing | 1068 | 1221 |
| LMArena Multi-Turn | 1062 | 1272 |
| WildBench | — | 80.2% |
Frequently asked questions
Is Mistral 7B better than Qwen2.5 72B Instruct?
Qwen2.5 72B Instruct is the stronger model overall, scoring 31.5 to 22.9 on the Noometry Index. Mistral 7B costs 9.8× less per token, which makes it the better buy when Qwen2.5 72B Instruct's lead doesn't matter for your workload.
Which is cheaper, Mistral 7B or Qwen2.5 72B Instruct?
Mistral 7B is cheaper. It lists at $0.25 per million input tokens and $0.25 per million output tokens; Qwen2.5 72B Instruct lists at $1.40 and $5.60.
Is Mistral 7B or Qwen2.5 72B Instruct better for coding?
Qwen2.5 72B Instruct scores higher on coding benchmarks: 32.8 versus 25.9 in the Noometry coding category.
Which has the bigger context window?
Qwen2.5 72B Instruct does, with 131K tokens against 8K.
How many benchmarks do Mistral 7B and Qwen2.5 72B Instruct share?
30 benchmarks have published results for both models. Mistral 7B has 37 scored results on Noometry and Qwen2.5 72B Instruct has 43.