Model comparison
Ministral 8B vs Qwen2.5 72B Instruct
Qwen2.5 72B Instruct is the stronger model overall, scoring 31.5 to 27.8 on the Noometry Index. Ministral 8B costs 24× less per token, which makes it the better buy when Qwen2.5 72B Instruct's lead doesn't matter for your workload.
Last verified . 15 shared benchmarks.
Mistral AI
Alibaba (Qwen)Summary
- They share 15 benchmarks with published results for both. Ministral 8B scores higher in 2 categories and Qwen2.5 72B Instruct in 7 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Qwen2.5 72B Instruct leads 26.9 to 12.5.
- The biggest single-benchmark swing is MATH Level 5: 14.9% for Ministral 8B and 63.2% for Qwen2.5 72B Instruct.
- Ministral 8B is cheaper at $0.10 / $0.10 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 128K.
Side by side
| Ministral 8B | Qwen2.5 72B Instruct | |
|---|---|---|
| Provider | Mistral AI | Alibaba (Qwen) |
| Noometry Index | 27.8 | 31.5 |
| Released | 2024-10-01 | 2024-09 |
| Weights | Open | Open |
| Context window | 128K | 131K |
| Max output | 128K | 8K |
| Input $ / M tokens | $0.10 | $1.40 |
| Output $ / M tokens | $0.10 | $5.60 |
| Results tracked | 17 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Ministral 8B leads
Ministral 8B: 34.4 (#226), Qwen2.5 72B Instruct: 32.8 (#256)
| Benchmark | Ministral 8B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Coding | 1202 | 1293 |
| WeirdML | — | 16% |
| BigCodeBench Instruct | — | 45.8% |
| BigCodeBench Complete | — | 55.9% |
Agentic & Tool Use Qwen2.5 72B Instruct leads
Ministral 8B: 16.2 (#140), Qwen2.5 72B Instruct: 22.0 (#124)
| Benchmark | Ministral 8B | Qwen2.5 72B Instruct |
|---|---|---|
| Berkeley Function Calling Leaderboard | 11.1% | — |
| TheAgentCompany | — | 5.7% |
| BALROG | — | 16.2% |
| METR Time Horizons | — | 35.8% |
Reasoning Qwen2.5 72B Instruct leads
Ministral 8B: 17.7 (#279), Qwen2.5 72B Instruct: 21.4 (#197)
| Benchmark | Ministral 8B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Hard Prompts | 1191 | 1271 |
| DTBench | 45.7% | 62.9% |
| LMCA | — | 13.4% |
| BIG-Bench Hard | — | 79.8% |
| Epoch Capabilities Index | — | 129.01 |
| ForecastBench | — | 57.5 |
| HellaSwag | — | 84.8% |
| PIQA | — | 82.6% |
| WinoGrande | — | 82.3% |
Math Ministral 8B leads
Ministral 8B: 23.7 (#262), Qwen2.5 72B Instruct: 18.1 (#281)
| Benchmark | Ministral 8B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Math | 1188 | 1283 |
| MATH Level 5 | 14.9% | 63.2% |
| OTIS Mock AIME 2024-2025 | — | 8.1% |
| Omni-MATH | — | 33% |
Knowledge Qwen2.5 72B Instruct leads
Ministral 8B: 12.5 (#292), Qwen2.5 72B Instruct: 26.9 (#247)
| Benchmark | Ministral 8B | Qwen2.5 72B Instruct |
|---|---|---|
| GPQA Diamond | 27.1% | 49.1% |
| LMArena Expert | 1170 | 1245 |
| MMLU-Pro | — | 63.1% |
| Confabulations | — | 19.1% |
| Vectara Hallucination Rate | 7.4% | — |
| GPQA (HELM) | — | 42.6% |
| ARC (AI2) Challenge | — | 94.5% |
| MMLU | — | 85.3% |
| TriviaQA | — | 71.9% |
Multilingual Qwen2.5 72B Instruct leads
Ministral 8B: 35.3 (#242), Qwen2.5 72B Instruct: 41.2 (#208)
| Benchmark | Ministral 8B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Non-English | 1165 | 1252 |
| LMArena Chinese | 1193 | 1272 |
| LMArena Russian | 1195 | 1264 |
| LMArena French | — | 1279 |
| LMArena German | — | 1234 |
| LMArena Japanese | — | 1180 |
| LMArena Korean | — | 1188 |
| LMArena Spanish | — | 1255 |
Instruction Following Qwen2.5 72B Instruct leads
Ministral 8B: 60.7 (#245), Qwen2.5 72B Instruct: 65.7 (#216)
| Benchmark | Ministral 8B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Instruction Following | 1161 | 1254 |
| IFEval | — | 80.6% |
Long Context Qwen2.5 72B Instruct leads
Ministral 8B: 36.7 (#222), Qwen2.5 72B Instruct: 38.9 (#183)
| Benchmark | Ministral 8B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Longer Query | 1212 | 1282 |
Writing & Preference Qwen2.5 72B Instruct leads
Ministral 8B: 39.8 (#241), Qwen2.5 72B Instruct: 46.9 (#210)
| Benchmark | Ministral 8B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Text | 1191 | 1269 |
| LMArena Creative Writing | 1175 | 1221 |
| LMArena Multi-Turn | 1166 | 1272 |
| WildBench | — | 80.2% |
Frequently asked questions
Is Ministral 8B better than Qwen2.5 72B Instruct?
Qwen2.5 72B Instruct is the stronger model overall, scoring 31.5 to 27.8 on the Noometry Index. Ministral 8B costs 24× 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, Ministral 8B or Qwen2.5 72B Instruct?
Ministral 8B is cheaper. It lists at $0.10 per million input tokens and $0.10 per million output tokens; Qwen2.5 72B Instruct lists at $1.40 and $5.60.
Is Ministral 8B or Qwen2.5 72B Instruct better for coding?
Ministral 8B scores higher on coding benchmarks: 34.4 versus 32.8 in the Noometry coding category.
Which has the bigger context window?
Qwen2.5 72B Instruct does, with 131K tokens against 128K.
How many benchmarks do Ministral 8B and Qwen2.5 72B Instruct share?
15 benchmarks have published results for both models. Ministral 8B has 17 scored results on Noometry and Qwen2.5 72B Instruct has 43.