Model comparison
Llama 3-70B vs Qwen2.5 72B Instruct
Qwen2.5 72B Instruct is the stronger model overall, scoring 31.5 to 28.5 on the Noometry Index.
Last verified . 27 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 27 benchmarks with published results for both. Llama 3-70B scores higher in 1 category and Qwen2.5 72B Instruct in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in multilingual, where Qwen2.5 72B Instruct leads 41.2 to 33.8.
- The biggest single-benchmark swing is MATH Level 5: 22.6% for Llama 3-70B and 63.2% for Qwen2.5 72B Instruct.
Side by side
| Llama 3-70B | Qwen2.5 72B Instruct | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 28.5 | 31.5 |
| Released | 2024-04-18 | 2024-09 |
| Weights | Open | Open |
| Context window | — | 131K |
| Max output | — | 8K |
| Input $ / M tokens | — | $1.40 |
| Output $ / M tokens | — | $5.60 |
| Results tracked | 31 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 3-70B leads
Llama 3-70B: 35.3 (#215), Qwen2.5 72B Instruct: 32.8 (#256)
| Benchmark | Llama 3-70B | Qwen2.5 72B Instruct |
|---|---|---|
| BigCodeBench Instruct | 43.6% | 45.8% |
| LMArena Coding | 1206 | 1293 |
| BigCodeBench Complete | 54.5% | 55.9% |
| WeirdML | — | 16% |
| HumanEval+ | 72% | — |
| MBPP+ | 69% | — |
Agentic & Tool Use Qwen2.5 72B Instruct leads
Llama 3-70B: 20.9 (#132), Qwen2.5 72B Instruct: 22.0 (#124)
| Benchmark | Llama 3-70B | Qwen2.5 72B Instruct |
|---|---|---|
| TheAgentCompany | — | 5.7% |
| Cybench | 5% | — |
| BALROG | — | 16.2% |
| METR Time Horizons | — | 35.8% |
Reasoning Qwen2.5 72B Instruct leads
Llama 3-70B: 17.4 (#284), Qwen2.5 72B Instruct: 21.4 (#197)
| Benchmark | Llama 3-70B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Hard Prompts | 1195 | 1271 |
| DTBench | 54.2% | 62.9% |
| Epoch Capabilities Index | 122.91 | 129.01 |
| ForecastBench | 57.1 | 57.5 |
| WinoGrande | 83.5% | 82.3% |
| Kagi LLM Benchmark | 35.1% | — |
| LMCA | — | 13.4% |
| BIG-Bench Hard | — | 79.8% |
| HellaSwag | — | 84.8% |
| PIQA | — | 82.6% |
Math Qwen2.5 72B Instruct leads
Llama 3-70B: 12.0 (#299), Qwen2.5 72B Instruct: 18.1 (#281)
| Benchmark | Llama 3-70B | Qwen2.5 72B Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 4.3% | 8.1% |
| LMArena Math | 1218 | 1283 |
| MATH Level 5 | 22.6% | 63.2% |
| Omni-MATH | — | 33% |
Knowledge Qwen2.5 72B Instruct leads
Llama 3-70B: 20.8 (#272), Qwen2.5 72B Instruct: 26.9 (#247)
| Benchmark | Llama 3-70B | Qwen2.5 72B Instruct |
|---|---|---|
| GPQA Diamond | 40.6% | 49.1% |
| LMArena Expert | 1149 | 1245 |
| MMLU | 79.3% | 85.3% |
| MMLU-Pro | — | 63.1% |
| Confabulations | — | 19.1% |
| GPQA (HELM) | — | 42.6% |
| ARC (AI2) Challenge | — | 94.5% |
| TriviaQA | — | 71.9% |
Multilingual Qwen2.5 72B Instruct leads
Llama 3-70B: 33.8 (#246), Qwen2.5 72B Instruct: 41.2 (#208)
| Benchmark | Llama 3-70B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Non-English | 1142 | 1252 |
| LMArena Chinese | 1114 | 1272 |
| LMArena French | 1231 | 1279 |
| LMArena German | 1169 | 1234 |
| LMArena Japanese | 1017 | 1180 |
| LMArena Korean | 1017 | 1188 |
| LMArena Russian | 1159 | 1264 |
| LMArena Spanish | 1241 | 1255 |
Instruction Following Qwen2.5 72B Instruct leads
Llama 3-70B: 62.8 (#233), Qwen2.5 72B Instruct: 65.7 (#216)
| Benchmark | Llama 3-70B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Instruction Following | 1194 | 1254 |
| IFEval | — | 80.6% |
Long Context Qwen2.5 72B Instruct leads
Llama 3-70B: 35.6 (#235), Qwen2.5 72B Instruct: 38.9 (#183)
| Benchmark | Llama 3-70B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Longer Query | 1174 | 1282 |
Writing & Preference Qwen2.5 72B Instruct leads
Llama 3-70B: 43.0 (#225), Qwen2.5 72B Instruct: 46.9 (#210)
| Benchmark | Llama 3-70B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Text | 1221 | 1269 |
| LMArena Creative Writing | 1210 | 1221 |
| LMArena Multi-Turn | 1223 | 1272 |
| WildBench | — | 80.2% |
Frequently asked questions
Is Llama 3-70B better than Qwen2.5 72B Instruct?
Qwen2.5 72B Instruct is the stronger model overall, scoring 31.5 to 28.5 on the Noometry Index.
Is Llama 3-70B or Qwen2.5 72B Instruct better for coding?
Llama 3-70B scores higher on coding benchmarks: 35.3 versus 32.8 in the Noometry coding category.
How many benchmarks do Llama 3-70B and Qwen2.5 72B Instruct share?
27 benchmarks have published results for both models. Llama 3-70B has 31 scored results on Noometry and Qwen2.5 72B Instruct has 43.