Model comparison
Llama 3-8B vs Qwen1.5-72B
Qwen1.5-72B is the stronger model overall, scoring 30.3 to 25.3 on the Noometry Index.
Last verified . 22 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 22 benchmarks with published results for both. Llama 3-8B scores higher in 1 category and Qwen1.5-72B in 7 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen1.5-72B leads 30.7 to 8.3.
Side by side
| Llama 3-8B | Qwen1.5-72B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 25.3 | 30.3 |
| Released | 2024-04-18 | 2024-02-04 |
| Weights | Open | Open |
| Context window | 8K | — |
| Max output | 8K | — |
| Input $ / M tokens | $0.14 | — |
| Output $ / M tokens | $0.14 | — |
| Results tracked | 34 | 22 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Llama 3-8B: 30.5 (#285), Qwen1.5-72B: 31.4 (#272)
| Benchmark | Llama 3-8B | Qwen1.5-72B |
|---|---|---|
| BigCodeBench Instruct | 31.9% | 33.2% |
| LMArena Coding | 1152 | 1166 |
| BigCodeBench Complete | 36.9% | 40.3% |
| HumanEval+ | 56.7% | 59.1% |
| MBPP+ | 54.8% | 61.6% |
Reasoning Qwen1.5-72B leads
Llama 3-8B: 13.8 (#321), Qwen1.5-72B: 21.2 (#201)
| Benchmark | Llama 3-8B | Qwen1.5-72B |
|---|---|---|
| LMArena Hard Prompts | 1133 | 1148 |
| Chess Puzzles | 0% | — |
| DTBench | 43.9% | — |
| Adversarial NLI | 57.3% | — |
| Epoch Capabilities Index | 116.33 | — |
| ForecastBench | 58.6 | — |
| WinoGrande | 75.7% | — |
Math Qwen1.5-72B leads
Llama 3-8B: 8.3 (#317), Qwen1.5-72B: 30.7 (#205)
| Benchmark | Llama 3-8B | Qwen1.5-72B |
|---|---|---|
| LMArena Math | 1151 | 1164 |
| OTIS Mock AIME 2024-2025 | 1.9% | — |
| MATH Level 5 | 6.1% | — |
Knowledge Qwen1.5-72B leads
Llama 3-8B: 7.8 (#303), Qwen1.5-72B: 11.5 (#295)
| Benchmark | Llama 3-8B | Qwen1.5-72B |
|---|---|---|
| GPQA Diamond | 26.1% | 28.8% |
| LMArena Expert | 1113 | 1136 |
| ARC (AI2) Challenge | 82.8% | — |
| MMLU | 68.8% | — |
| OpenBookQA | 82.6% | — |
| TriviaQA | 67.7% | — |
Multilingual Qwen1.5-72B leads
Llama 3-8B: 31.0 (#256), Qwen1.5-72B: 33.4 (#248)
| Benchmark | Llama 3-8B | Qwen1.5-72B |
|---|---|---|
| LMArena Non-English | 1098 | 1135 |
| LMArena Chinese | 1076 | 1186 |
| LMArena French | 1158 | 1158 |
| LMArena German | 1104 | 1084 |
| LMArena Japanese | 967 | 1061 |
| LMArena Korean | 1003 | 1049 |
| LMArena Russian | 1109 | 1104 |
| LMArena Spanish | 1172 | 1110 |
Instruction Following Too close to call
Llama 3-8B: 58.7 (#255), Qwen1.5-72B: 59.6 (#251)
| Benchmark | Llama 3-8B | Qwen1.5-72B |
|---|---|---|
| LMArena Instruction Following | 1127 | 1142 |
Long Context Too close to call
Llama 3-8B: 34.2 (#246), Qwen1.5-72B: 35.0 (#238)
| Benchmark | Llama 3-8B | Qwen1.5-72B |
|---|---|---|
| LMArena Longer Query | 1128 | 1157 |
Writing & Preference Too close to call
Llama 3-8B: 37.7 (#251), Qwen1.5-72B: 37.4 (#253)
| Benchmark | Llama 3-8B | Qwen1.5-72B |
|---|---|---|
| LMArena Text | 1166 | 1166 |
| LMArena Creative Writing | 1150 | 1137 |
| LMArena Multi-Turn | 1152 | 1160 |
Frequently asked questions
Is Llama 3-8B better than Qwen1.5-72B?
Qwen1.5-72B is the stronger model overall, scoring 30.3 to 25.3 on the Noometry Index.
Is Llama 3-8B or Qwen1.5-72B better for coding?
They score almost the same on coding (30.5 vs 31.4); test both on your own repository before choosing.
How many benchmarks do Llama 3-8B and Qwen1.5-72B share?
22 benchmarks have published results for both models. Llama 3-8B has 34 scored results on Noometry and Qwen1.5-72B has 22.