Model comparison
Qwen2-72B vs Qwen3 32B
Qwen3 32B is the stronger model overall, scoring 39.2 to 30.0 on the Noometry Index.
Last verified . 15 shared benchmarks.
Alibaba (Qwen)Summary
- They share 15 benchmarks with published results for both. Qwen2-72B scores higher in 1 category and Qwen3 32B in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Qwen3 32B leads 40.0 to 21.2.
- The biggest single-benchmark swing is GPQA Diamond: 40.8% for Qwen2-72B and 65.7% for Qwen3 32B.
Side by side
| Qwen2-72B | Qwen3 32B | |
|---|---|---|
| Provider | Alibaba (Qwen) | Alibaba (Qwen) |
| Noometry Index | 30.0 | 39.2 |
| Released | 2024-06-07 | 2025-04 |
| Weights | Open | Open |
| Context window | — | 131K |
| Max output | — | 16K |
| Input $ / M tokens | — | $0.70 |
| Output $ / M tokens | — | $2.80 |
| Results tracked | 26 | 26 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3 32B leads
Qwen2-72B: 29.1 (#310), Qwen3 32B: 37.7 (#190)
| Benchmark | Qwen2-72B | Qwen3 32B |
|---|---|---|
| LMArena Coding | 1196 | 1358 |
| Aider Polyglot | — | 40% |
| SciCode | — | 35.4% |
| WeirdML | 11.3% | — |
| BigCodeBench Instruct | 38.5% | — |
| BigCodeBench Complete | 54% | — |
Agentic & Tool Use Qwen3 32B leads
Qwen2-72B: 17.0 (#146), Qwen3 32B: 32.6 (#62)
| Benchmark | Qwen2-72B | Qwen3 32B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 48.7% |
| TheAgentCompany | 1.1% | — |
| METR Time Horizons | 29.9% | — |
Reasoning Qwen2-72B leads
Qwen2-72B: 23.2 (#181), Qwen3 32B: 20.2 (#241)
| Benchmark | Qwen2-72B | Qwen3 32B |
|---|---|---|
| LMArena Hard Prompts | 1191 | 1334 |
| Epoch Capabilities Index | 125.28 | 138.51 |
| Kagi LLM Benchmark | — | 54.9% |
| CritPt | — | 0.3% |
| Chess Puzzles | — | 5% |
| DTBench | — | 67.5% |
| LMCA | — | 17.3% |
Math Qwen3 32B leads
Qwen2-72B: 30.2 (#236), Qwen3 32B: 39.7 (#99)
| Benchmark | Qwen2-72B | Qwen3 32B |
|---|---|---|
| LMArena Math | 1235 | 1399 |
| OTIS Mock AIME 2024-2025 | — | 66.9% |
| MATH Level 5 | 39.1% | — |
Knowledge Qwen3 32B leads
Qwen2-72B: 21.2 (#275), Qwen3 32B: 40.0 (#125)
| Benchmark | Qwen2-72B | Qwen3 32B |
|---|---|---|
| GPQA Diamond | 40.8% | 65.7% |
| LMArena Expert | 1171 | 1362 |
| Vectara Hallucination Rate | — | 5.9% |
| MMLU | 82.4% | — |
Multilingual Qwen3 32B leads
Qwen2-72B: 35.9 (#244), Qwen3 32B: 45.6 (#167)
| Benchmark | Qwen2-72B | Qwen3 32B |
|---|---|---|
| LMArena Non-English | 1176 | 1317 |
| LMArena Chinese | 1240 | 1357 |
| LMArena German | 1151 | 1341 |
| LMArena Russian | 1169 | 1311 |
| LMArena French | 1170 | — |
| LMArena Japanese | 1111 | — |
| LMArena Korean | 1083 | — |
| LMArena Spanish | 1169 | — |
Instruction Following Qwen3 32B leads
Qwen2-72B: 61.7 (#241), Qwen3 32B: 68.9 (#179)
| Benchmark | Qwen2-72B | Qwen3 32B |
|---|---|---|
| LMArena Instruction Following | 1181 | 1305 |
Long Context Qwen3 32B leads
Qwen2-72B: 36.1 (#235), Qwen3 32B: 43.8 (#87)
| Benchmark | Qwen2-72B | Qwen3 32B |
|---|---|---|
| LMArena Longer Query | 1192 | 1327 |
| Fiction.LiveBench | — | 74.2% |
Writing & Preference Qwen3 32B leads
Qwen2-72B: 40.8 (#241), Qwen3 32B: 52.9 (#163)
| Benchmark | Qwen2-72B | Qwen3 32B |
|---|---|---|
| LMArena Text | 1203 | 1340 |
| LMArena Creative Writing | 1181 | 1297 |
| LMArena Multi-Turn | 1196 | 1331 |
Frequently asked questions
Is Qwen2-72B better than Qwen3 32B?
Qwen3 32B is the stronger model overall, scoring 39.2 to 30.0 on the Noometry Index.
Is Qwen2-72B or Qwen3 32B better for coding?
Qwen3 32B scores higher on coding benchmarks: 37.7 versus 29.1 in the Noometry coding category.
How many benchmarks do Qwen2-72B and Qwen3 32B share?
15 benchmarks have published results for both models. Qwen2-72B has 26 scored results on Noometry and Qwen3 32B has 26.