Model comparison
Llama 3-70B vs Qwen2-72B
Qwen2-72B is the stronger model overall, scoring 30.0 to 28.8 on the Noometry Index.
Last verified . 23 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 23 benchmarks with published results for both. Llama 3-70B scores higher in 4 categories and Qwen2-72B in 5 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen2-72B leads 30.2 to 12.8.
- The biggest single-benchmark swing is MATH Level 5: 22.6% for Llama 3-70B and 39.1% for Qwen2-72B.
Side by side
| Llama 3-70B | Qwen2-72B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 28.8 | 30.0 |
| Released | 2024-04-18 | 2024-06-07 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 31 | 26 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 3-70B leads
Llama 3-70B: 35.8 (#218), Qwen2-72B: 29.1 (#310)
| Benchmark | Llama 3-70B | Qwen2-72B |
|---|---|---|
| BigCodeBench Instruct | 43.6% | 38.5% |
| LMArena Coding | 1206 | 1196 |
| BigCodeBench Complete | 54.5% | 54% |
| WeirdML | — | 11.3% |
| HumanEval+ | 72% | — |
| MBPP+ | 69% | — |
Agentic & Tool Use Llama 3-70B leads
Llama 3-70B: 21.1 (#139), Qwen2-72B: 17.0 (#146)
| Benchmark | Llama 3-70B | Qwen2-72B |
|---|---|---|
| TheAgentCompany | — | 1.1% |
| Cybench | 5% | — |
| METR Time Horizons | — | 29.9% |
Reasoning Qwen2-72B leads
Llama 3-70B: 18.0 (#288), Qwen2-72B: 23.2 (#181)
| Benchmark | Llama 3-70B | Qwen2-72B |
|---|---|---|
| LMArena Hard Prompts | 1195 | 1191 |
| Epoch Capabilities Index | 122.93 | 125.28 |
| Kagi LLM Benchmark | 35.1% | — |
| DTBench | 54.2% | — |
| ForecastBench | 57.1 | — |
| WinoGrande | 83.5% | — |
Math Qwen2-72B leads
Llama 3-70B: 12.8 (#305), Qwen2-72B: 30.2 (#236)
| Benchmark | Llama 3-70B | Qwen2-72B |
|---|---|---|
| LMArena Math | 1218 | 1235 |
| MATH Level 5 | 22.6% | 39.1% |
| OTIS Mock AIME 2024-2025 | 4.3% | — |
Knowledge Too close to call
Llama 3-70B: 20.8 (#277), Qwen2-72B: 21.2 (#275)
| Benchmark | Llama 3-70B | Qwen2-72B |
|---|---|---|
| GPQA Diamond | 40.6% | 40.8% |
| LMArena Expert | 1149 | 1171 |
| MMLU | 79.3% | 82.4% |
Multilingual Qwen2-72B leads
Llama 3-70B: 33.6 (#251), Qwen2-72B: 35.9 (#244)
| Benchmark | Llama 3-70B | Qwen2-72B |
|---|---|---|
| LMArena Non-English | 1142 | 1176 |
| LMArena Chinese | 1114 | 1240 |
| LMArena French | 1232 | 1170 |
| LMArena German | 1169 | 1151 |
| LMArena Japanese | 1017 | 1111 |
| LMArena Korean | 1017 | 1083 |
| LMArena Russian | 1159 | 1169 |
| LMArena Spanish | 1241 | 1169 |
Instruction Following Too close to call
Llama 3-70B: 62.5 (#238), Qwen2-72B: 61.7 (#241)
| Benchmark | Llama 3-70B | Qwen2-72B |
|---|---|---|
| LMArena Instruction Following | 1194 | 1181 |
Long Context Too close to call
Llama 3-70B: 35.6 (#240), Qwen2-72B: 36.1 (#235)
| Benchmark | Llama 3-70B | Qwen2-72B |
|---|---|---|
| LMArena Longer Query | 1174 | 1192 |
Writing & Preference Llama 3-70B leads
Llama 3-70B: 42.8 (#231), Qwen2-72B: 40.8 (#241)
| Benchmark | Llama 3-70B | Qwen2-72B |
|---|---|---|
| LMArena Text | 1221 | 1203 |
| LMArena Creative Writing | 1210 | 1181 |
| LMArena Multi-Turn | 1223 | 1196 |
Frequently asked questions
Is Llama 3-70B better than Qwen2-72B?
Qwen2-72B is the stronger model overall, scoring 30.0 to 28.8 on the Noometry Index.
Is Llama 3-70B or Qwen2-72B better for coding?
Llama 3-70B scores higher on coding benchmarks: 35.8 versus 29.1 in the Noometry coding category.
How many benchmarks do Llama 3-70B and Qwen2-72B share?
23 benchmarks have published results for both models. Llama 3-70B has 31 scored results on Noometry and Qwen2-72B has 26.