Model comparison
Mistral vs Qwen3-4B
Qwen3-4B is the stronger model overall, scoring 31.9 to 29.9 on the Noometry Index.
Last verified . 0 shared benchmarks.
Mistral AI
Alibaba (Qwen)Summary
- The widest gap is in knowledge, where Qwen3-4B leads 33.0 to 16.6.
- Qwen3-4B has downloadable open weights; the other is API-only.
Side by side
| Mistral | Qwen3-4B | |
|---|---|---|
| Provider | Mistral AI | Alibaba (Qwen) |
| Noometry Index | 29.9 | 31.9 |
| Released | — | 2025-04-29 |
| Weights | Proprietary | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 22 | 6 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Not comparable
Mistral: 33.8 (#250), Qwen3-4B: —
| Benchmark | Mistral | Qwen3-4B |
|---|---|---|
| LMArena Coding | 1162 | — |
Agentic & Tool Use Not comparable
Mistral: —, Qwen3-4B: 27.6 (#100)
| Benchmark | Mistral | Qwen3-4B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 35.7% |
Reasoning Mistral leads
Mistral: 22.2 (#200), Qwen3-4B: 19.2 (#268)
| Benchmark | Mistral | Qwen3-4B |
|---|---|---|
| Chess Puzzles | — | 4% |
| LMArena Hard Prompts | 1149 | — |
Math Qwen3-4B leads
Mistral: 22.3 (#278), Qwen3-4B: 29.7 (#240)
| Benchmark | Mistral | Qwen3-4B |
|---|---|---|
| MathArena Final-Answer Competitions | — | 38.5% |
| OTIS Mock AIME 2024-2025 | — | 52.2% |
| Omni-MATH | 7.2% | — |
| LMArena Math | 1180 | — |
Knowledge Qwen3-4B leads
Mistral: 16.6 (#288), Qwen3-4B: 33.0 (#208)
| Benchmark | Mistral | Qwen3-4B |
|---|---|---|
| GPQA Diamond | — | 52.3% |
| MMLU-Pro | 27.7% | — |
| Vectara Hallucination Rate | — | 5.7% |
| GPQA (HELM) | 30.3% | — |
| LMArena Expert | 1125 | — |
Multilingual Not comparable
Mistral: 32.8 (#254), Qwen3-4B: —
| Benchmark | Mistral | Qwen3-4B |
|---|---|---|
| LMArena Non-English | 1129 | — |
| LMArena Chinese | 1109 | — |
| LMArena French | 1180 | — |
| LMArena German | 1155 | — |
| LMArena Japanese | 1013 | — |
| LMArena Korean | 1032 | — |
| LMArena Russian | 1168 | — |
| LMArena Spanish | 1143 | — |
Instruction Following Not comparable
Mistral: 52.6 (#288), Qwen3-4B: —
| Benchmark | Mistral | Qwen3-4B |
|---|---|---|
| IFEval | 56.8% | — |
| LMArena Instruction Following | 1152 | — |
Long Context Not comparable
Mistral: 35.0 (#245), Qwen3-4B: —
| Benchmark | Mistral | Qwen3-4B |
|---|---|---|
| LMArena Longer Query | 1153 | — |
Writing & Preference Not comparable
Mistral: 37.0 (#260), Qwen3-4B: —
| Benchmark | Mistral | Qwen3-4B |
|---|---|---|
| LMArena Text | 1165 | — |
| LMArena Creative Writing | 1158 | — |
| WildBench | 66% | — |
| LMArena Multi-Turn | 1147 | — |
Frequently asked questions
Is Mistral better than Qwen3-4B?
Qwen3-4B is the stronger model overall, scoring 31.9 to 29.9 on the Noometry Index.
How many benchmarks do Mistral and Qwen3-4B share?
0 benchmarks have published results for both models. Mistral has 22 scored results on Noometry and Qwen3-4B has 6.