Model comparison
Llama 2-70B vs Qwen3 32B
Qwen3 32B is the stronger model overall, scoring 39.2 to 24.4 on the Noometry Index.
Last verified . 17 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 17 benchmarks with published results for both. Llama 2-70B scores higher in 0 categories and Qwen3 32B in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Qwen3 32B leads 40.0 to 7.4.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 0% for Llama 2-70B and 66.9% for Qwen3 32B.
Side by side
| Llama 2-70B | Qwen3 32B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 24.4 | 39.2 |
| Released | 2023-07-18 | 2025-04 |
| Weights | Open | Open |
| Context window | — | 131K |
| Max output | — | 16K |
| Input $ / M tokens | — | $0.70 |
| Output $ / M tokens | — | $2.80 |
| Results tracked | 35 | 26 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3 32B leads
Llama 2-70B: 31.4 (#286), Qwen3 32B: 37.7 (#190)
| Benchmark | Llama 2-70B | Qwen3 32B |
|---|---|---|
| LMArena Coding | 1079 | 1358 |
| Aider Polyglot | — | 40% |
| SciCode | — | 35.4% |
Agentic & Tool Use Not comparable
Llama 2-70B: —, Qwen3 32B: 32.6 (#62)
| Benchmark | Llama 2-70B | Qwen3 32B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 48.7% |
Reasoning Qwen3 32B leads
Llama 2-70B: 14.4 (#325), Qwen3 32B: 20.2 (#241)
| Benchmark | Llama 2-70B | Qwen3 32B |
|---|---|---|
| LMArena Hard Prompts | 1073 | 1334 |
| DTBench | 41.6% | 67.5% |
| Epoch Capabilities Index | 113.79 | 138.51 |
| Kagi LLM Benchmark | — | 54.9% |
| CritPt | — | 0.3% |
| Chess Puzzles | — | 5% |
| LMCA | — | 17.3% |
| 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 Qwen3 32B leads
Llama 2-70B: 8.1 (#326), Qwen3 32B: 39.7 (#99)
| Benchmark | Llama 2-70B | Qwen3 32B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 0% | 66.9% |
| LMArena Math | 1091 | 1399 |
| MATH Level 5 | 3.3% | — |
| GSM8K | 69.6% | — |
Knowledge Qwen3 32B leads
Llama 2-70B: 7.4 (#310), Qwen3 32B: 40.0 (#125)
| Benchmark | Llama 2-70B | Qwen3 32B |
|---|---|---|
| GPQA Diamond | 26.3% | 65.7% |
| LMArena Expert | 1039 | 1362 |
| Vectara Hallucination Rate | — | 5.9% |
| ARC (AI2) Challenge | 78.3% | — |
| BoolQ | 88.6% | — |
| MMLU | 69.9% | — |
| OpenBookQA | 60.2% | — |
| TriviaQA | 87.6% | — |
Multilingual Qwen3 32B leads
Llama 2-70B: 27.7 (#274), Qwen3 32B: 45.6 (#167)
| Benchmark | Llama 2-70B | Qwen3 32B |
|---|---|---|
| LMArena Non-English | 1045 | 1317 |
| LMArena Chinese | 995 | 1357 |
| LMArena German | 1041 | 1341 |
| LMArena Russian | 1083 | 1311 |
| LMArena French | 1090 | — |
| LMArena Japanese | 927 | — |
| LMArena Korean | 964 | — |
| LMArena Spanish | 1143 | — |
Instruction Following Qwen3 32B leads
Llama 2-70B: 54.9 (#278), Qwen3 32B: 68.9 (#179)
| Benchmark | Llama 2-70B | Qwen3 32B |
|---|---|---|
| LMArena Instruction Following | 1071 | 1305 |
Long Context Qwen3 32B leads
Llama 2-70B: 32.3 (#270), Qwen3 32B: 43.8 (#87)
| Benchmark | Llama 2-70B | Qwen3 32B |
|---|---|---|
| LMArena Longer Query | 1062 | 1327 |
| Fiction.LiveBench | — | 74.2% |
Writing & Preference Qwen3 32B leads
Llama 2-70B: 32.3 (#279), Qwen3 32B: 52.9 (#163)
| Benchmark | Llama 2-70B | Qwen3 32B |
|---|---|---|
| LMArena Text | 1115 | 1340 |
| LMArena Creative Writing | 1075 | 1297 |
| LMArena Multi-Turn | 1088 | 1331 |
Frequently asked questions
Is Llama 2-70B better than Qwen3 32B?
Qwen3 32B is the stronger model overall, scoring 39.2 to 24.4 on the Noometry Index.
Is Llama 2-70B or Qwen3 32B better for coding?
Qwen3 32B scores higher on coding benchmarks: 37.7 versus 31.4 in the Noometry coding category.
How many benchmarks do Llama 2-70B and Qwen3 32B share?
17 benchmarks have published results for both models. Llama 2-70B has 35 scored results on Noometry and Qwen3 32B has 26.