Model comparison
Llama 3.1-405B vs Qwen3-4B
Qwen3-4B is the stronger model overall, scoring 31.5 to 30.4 on the Noometry Index.
Last verified . 2 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 2 benchmarks with published results for both. Llama 3.1-405B scores higher in 0 categories and Qwen3-4B in 4 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3-4B leads 28.5 to 17.3.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 9.7% for Llama 3.1-405B and 52.2% for Qwen3-4B.
Side by side
| Llama 3.1-405B | Qwen3-4B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 30.4 | 31.5 |
| Released | 2024-07-23 | 2025-04-29 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 42 | 6 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Not comparable
Llama 3.1-405B: 32.7 (#257), Qwen3-4B: —
| Benchmark | Llama 3.1-405B | Qwen3-4B |
|---|---|---|
| WeirdML | 21.4% | — |
| LMArena Coding | 1291 | — |
Agentic & Tool Use Qwen3-4B leads
Llama 3.1-405B: 20.9 (#131), Qwen3-4B: 27.3 (#94)
| Benchmark | Llama 3.1-405B | Qwen3-4B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 35.7% |
| TheAgentCompany | 7.4% | — |
| Cybench | 7.5% | — |
Reasoning Qwen3-4B leads
Llama 3.1-405B: 16.4 (#293), Qwen3-4B: 18.5 (#264)
| Benchmark | Llama 3.1-405B | Qwen3-4B |
|---|---|---|
| SimpleBench | 23% | — |
| Kagi LLM Benchmark | 45% | — |
| Chess Puzzles | — | 4% |
| LMArena Hard Prompts | 1270 | — |
| DTBench | 61.4% | — |
| BIG-Bench Hard | 82.9% | — |
| Epoch Capabilities Index | 128.76 | — |
| ForecastBench | 59.9 | — |
| HellaSwag | 89.2% | — |
| PIQA | 85.9% | — |
| WinoGrande | 89.2% | — |
Math Qwen3-4B leads
Llama 3.1-405B: 17.3 (#283), Qwen3-4B: 28.5 (#228)
| Benchmark | Llama 3.1-405B | Qwen3-4B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 9.7% | 52.2% |
| MathArena Final-Answer Competitions | — | 38.5% |
| Omni-MATH | 24.9% | — |
| LMArena Math | 1281 | — |
| MATH Level 5 | 49.8% | — |
Knowledge Qwen3-4B leads
Llama 3.1-405B: 30.4 (#221), Qwen3-4B: 33.0 (#198)
| Benchmark | Llama 3.1-405B | Qwen3-4B |
|---|---|---|
| GPQA Diamond | 50.9% | 52.3% |
| MMLU-Pro | 72.3% | — |
| Confabulations | 17.6% | — |
| Vectara Hallucination Rate | — | 5.7% |
| GPQA (HELM) | 52.2% | — |
| LMArena Expert | 1243 | — |
| ARC (AI2) Challenge | 95.3% | — |
| MMLU | 84.5% | — |
| TriviaQA | 82.7% | — |
Multilingual Not comparable
Llama 3.1-405B: 40.9 (#209), Qwen3-4B: —
| Benchmark | Llama 3.1-405B | Qwen3-4B |
|---|---|---|
| LMArena Non-English | 1248 | — |
| LMArena Chinese | 1242 | — |
| LMArena French | 1278 | — |
| LMArena German | 1252 | — |
| LMArena Japanese | 1208 | — |
| LMArena Korean | 1184 | — |
| LMArena Russian | 1265 | — |
| LMArena Spanish | 1259 | — |
Instruction Following Not comparable
Llama 3.1-405B: 66.1 (#209), Qwen3-4B: —
| Benchmark | Llama 3.1-405B | Qwen3-4B |
|---|---|---|
| IFEval | 81.1% | — |
| LMArena Instruction Following | 1259 | — |
Long Context Not comparable
Llama 3.1-405B: 38.4 (#192), Qwen3-4B: —
| Benchmark | Llama 3.1-405B | Qwen3-4B |
|---|---|---|
| LMArena Longer Query | 1266 | — |
Writing & Preference Not comparable
Llama 3.1-405B: 39.1 (#246), Qwen3-4B: —
| Benchmark | Llama 3.1-405B | Qwen3-4B |
|---|---|---|
| LMArena Text | 1284 | — |
| LMArena Creative Writing | 1262 | — |
| EQ-Bench Creative Writing | 870 | — |
| WildBench | 78.3% | — |
| LMArena Multi-Turn | 1297 | — |
Frequently asked questions
Is Llama 3.1-405B better than Qwen3-4B?
Qwen3-4B is the stronger model overall, scoring 31.5 to 30.4 on the Noometry Index.
How many benchmarks do Llama 3.1-405B and Qwen3-4B share?
2 benchmarks have published results for both models. Llama 3.1-405B has 42 scored results on Noometry and Qwen3-4B has 6.