Model comparison
Llama 3-8B vs Qwen2-72B
Qwen2-72B is the stronger model overall, scoring 29.5 to 25.3 on the Noometry Index.
Last verified . 23 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 23 benchmarks with published results for both. Llama 3-8B scores higher in 1 category and Qwen2-72B in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen2-72B leads 27.9 to 8.3.
- The biggest single-benchmark swing is MATH Level 5: 6.1% for Llama 3-8B and 39.1% for Qwen2-72B.
Side by side
| Llama 3-8B | Qwen2-72B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 25.3 | 29.5 |
| Released | 2024-04-18 | 2024-06-07 |
| Weights | Open | Open |
| Context window | 8K | — |
| Max output | 8K | — |
| Input $ / M tokens | $0.14 | — |
| Output $ / M tokens | $0.14 | — |
| Results tracked | 34 | 26 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 3-8B leads
Llama 3-8B: 30.5 (#285), Qwen2-72B: 28.7 (#303)
| Benchmark | Llama 3-8B | Qwen2-72B |
|---|---|---|
| BigCodeBench Instruct | 31.9% | 38.5% |
| LMArena Coding | 1152 | 1196 |
| BigCodeBench Complete | 36.9% | 54% |
| WeirdML | — | 11.3% |
| HumanEval+ | 56.7% | — |
| MBPP+ | 54.8% | — |
Agentic & Tool Use Not comparable
Llama 3-8B: —, Qwen2-72B: 16.9 (#138)
| Benchmark | Llama 3-8B | Qwen2-72B |
|---|---|---|
| TheAgentCompany | — | 1.1% |
| METR Time Horizons | — | 29.9% |
Reasoning Qwen2-72B leads
Llama 3-8B: 13.8 (#321), Qwen2-72B: 22.2 (#184)
| Benchmark | Llama 3-8B | Qwen2-72B |
|---|---|---|
| LMArena Hard Prompts | 1133 | 1191 |
| Epoch Capabilities Index | 116.33 | 125.27 |
| Chess Puzzles | 0% | — |
| DTBench | 43.9% | — |
| Adversarial NLI | 57.3% | — |
| ForecastBench | 58.6 | — |
| WinoGrande | 75.7% | — |
Math Qwen2-72B leads
Llama 3-8B: 8.3 (#317), Qwen2-72B: 27.9 (#232)
| Benchmark | Llama 3-8B | Qwen2-72B |
|---|---|---|
| LMArena Math | 1151 | 1235 |
| MATH Level 5 | 6.1% | 39.1% |
| OTIS Mock AIME 2024-2025 | 1.9% | — |
Knowledge Qwen2-72B leads
Llama 3-8B: 7.8 (#303), Qwen2-72B: 21.2 (#270)
| Benchmark | Llama 3-8B | Qwen2-72B |
|---|---|---|
| GPQA Diamond | 26.1% | 40.8% |
| LMArena Expert | 1113 | 1171 |
| MMLU | 68.8% | 82.4% |
| ARC (AI2) Challenge | 82.8% | — |
| OpenBookQA | 82.6% | — |
| TriviaQA | 67.7% | — |
Multilingual Qwen2-72B leads
Llama 3-8B: 31.0 (#256), Qwen2-72B: 36.0 (#239)
| Benchmark | Llama 3-8B | Qwen2-72B |
|---|---|---|
| LMArena Non-English | 1098 | 1176 |
| LMArena Chinese | 1076 | 1240 |
| LMArena French | 1158 | 1169 |
| LMArena German | 1104 | 1151 |
| LMArena Japanese | 967 | 1111 |
| LMArena Korean | 1003 | 1083 |
| LMArena Russian | 1109 | 1169 |
| LMArena Spanish | 1172 | 1168 |
Instruction Following Qwen2-72B leads
Llama 3-8B: 58.7 (#255), Qwen2-72B: 62.0 (#236)
| Benchmark | Llama 3-8B | Qwen2-72B |
|---|---|---|
| LMArena Instruction Following | 1127 | 1181 |
Long Context Qwen2-72B leads
Llama 3-8B: 34.2 (#246), Qwen2-72B: 36.1 (#230)
| Benchmark | Llama 3-8B | Qwen2-72B |
|---|---|---|
| LMArena Longer Query | 1128 | 1193 |
Writing & Preference Qwen2-72B leads
Llama 3-8B: 37.7 (#251), Qwen2-72B: 41.0 (#236)
| Benchmark | Llama 3-8B | Qwen2-72B |
|---|---|---|
| LMArena Text | 1166 | 1203 |
| LMArena Creative Writing | 1150 | 1181 |
| LMArena Multi-Turn | 1152 | 1195 |
Frequently asked questions
Is Llama 3-8B better than Qwen2-72B?
Qwen2-72B is the stronger model overall, scoring 29.5 to 25.3 on the Noometry Index.
Is Llama 3-8B or Qwen2-72B better for coding?
Llama 3-8B scores higher on coding benchmarks: 30.5 versus 28.7 in the Noometry coding category.
How many benchmarks do Llama 3-8B and Qwen2-72B share?
23 benchmarks have published results for both models. Llama 3-8B has 34 scored results on Noometry and Qwen2-72B has 26.