Model comparison
Llama 2-70B vs Qwen2-72B
Qwen2-72B is the stronger model overall, scoring 30.0 to 24.4 on the Noometry Index.
Last verified . 21 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 21 benchmarks with published results for both. Llama 2-70B 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 30.2 to 8.1.
- The biggest single-benchmark swing is MATH Level 5: 3.3% for Llama 2-70B and 39.1% for Qwen2-72B.
Side by side
| Llama 2-70B | Qwen2-72B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 24.4 | 30.0 |
| Released | 2023-07-18 | 2024-06-07 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 35 | 26 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 2-70B leads
Llama 2-70B: 31.4 (#286), Qwen2-72B: 29.1 (#310)
| Benchmark | Llama 2-70B | Qwen2-72B |
|---|---|---|
| LMArena Coding | 1079 | 1196 |
| WeirdML | — | 11.3% |
| BigCodeBench Instruct | — | 38.5% |
| BigCodeBench Complete | — | 54% |
Agentic & Tool Use Not comparable
Llama 2-70B: —, Qwen2-72B: 17.0 (#146)
| Benchmark | Llama 2-70B | Qwen2-72B |
|---|---|---|
| TheAgentCompany | — | 1.1% |
| METR Time Horizons | — | 29.9% |
Reasoning Qwen2-72B leads
Llama 2-70B: 14.4 (#325), Qwen2-72B: 23.2 (#181)
| Benchmark | Llama 2-70B | Qwen2-72B |
|---|---|---|
| LMArena Hard Prompts | 1073 | 1191 |
| Epoch Capabilities Index | 113.79 | 125.28 |
| DTBench | 41.6% | — |
| BIG-Bench Hard | 64.9% | — |
| CommonsenseQA 2.0 | 50% | — |
| ForecastBench | 51.4 | — |
| HellaSwag | 85.3% | — |
| LAMBADA | 78.9% | — |
| PIQA | 82.8% | — |
| WinoGrande | 80.2% | — |
Math Qwen2-72B leads
Llama 2-70B: 8.1 (#326), Qwen2-72B: 30.2 (#236)
| Benchmark | Llama 2-70B | Qwen2-72B |
|---|---|---|
| LMArena Math | 1091 | 1235 |
| MATH Level 5 | 3.3% | 39.1% |
| OTIS Mock AIME 2024-2025 | 0% | — |
| GSM8K | 69.6% | — |
Knowledge Qwen2-72B leads
Llama 2-70B: 7.4 (#310), Qwen2-72B: 21.2 (#275)
| Benchmark | Llama 2-70B | Qwen2-72B |
|---|---|---|
| GPQA Diamond | 26.3% | 40.8% |
| LMArena Expert | 1039 | 1171 |
| MMLU | 69.9% | 82.4% |
| ARC (AI2) Challenge | 78.3% | — |
| BoolQ | 88.6% | — |
| OpenBookQA | 60.2% | — |
| TriviaQA | 87.6% | — |
Multilingual Qwen2-72B leads
Llama 2-70B: 27.7 (#274), Qwen2-72B: 35.9 (#244)
| Benchmark | Llama 2-70B | Qwen2-72B |
|---|---|---|
| LMArena Non-English | 1045 | 1176 |
| LMArena Chinese | 995 | 1240 |
| LMArena French | 1090 | 1170 |
| LMArena German | 1041 | 1151 |
| LMArena Japanese | 927 | 1111 |
| LMArena Korean | 964 | 1083 |
| LMArena Russian | 1083 | 1169 |
| LMArena Spanish | 1143 | 1169 |
Instruction Following Qwen2-72B leads
Llama 2-70B: 54.9 (#278), Qwen2-72B: 61.7 (#241)
| Benchmark | Llama 2-70B | Qwen2-72B |
|---|---|---|
| LMArena Instruction Following | 1071 | 1181 |
Long Context Qwen2-72B leads
Llama 2-70B: 32.3 (#270), Qwen2-72B: 36.1 (#235)
| Benchmark | Llama 2-70B | Qwen2-72B |
|---|---|---|
| LMArena Longer Query | 1062 | 1192 |
Writing & Preference Qwen2-72B leads
Llama 2-70B: 32.3 (#279), Qwen2-72B: 40.8 (#241)
| Benchmark | Llama 2-70B | Qwen2-72B |
|---|---|---|
| LMArena Text | 1115 | 1203 |
| LMArena Creative Writing | 1075 | 1181 |
| LMArena Multi-Turn | 1088 | 1196 |
Frequently asked questions
Is Llama 2-70B better than Qwen2-72B?
Qwen2-72B is the stronger model overall, scoring 30.0 to 24.4 on the Noometry Index.
Is Llama 2-70B or Qwen2-72B better for coding?
Llama 2-70B scores higher on coding benchmarks: 31.4 versus 29.1 in the Noometry coding category.
How many benchmarks do Llama 2-70B and Qwen2-72B share?
21 benchmarks have published results for both models. Llama 2-70B has 35 scored results on Noometry and Qwen2-72B has 26.