Model comparison
Qwen2-72B vs Qwen3.8 27B
Qwen3.8 27B is the stronger model overall, scoring 46.0 to 30.0 on the Noometry Index.
Last verified . 18 shared benchmarks.
Alibaba (Qwen)Summary
- They share 18 benchmarks with published results for both. Qwen2-72B scores higher in 0 categories and Qwen3.8 27B in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Qwen3.8 27B leads 65.8 to 40.8.
Side by side
| Qwen2-72B | Qwen3.8 27B | |
|---|---|---|
| Provider | Alibaba (Qwen) | Alibaba (Qwen) |
| Noometry Index | 30.0 | 46.0 |
| Released | 2024-06-07 | 2026-08-14 |
| Weights | Open | Open |
| Context window | — | 262K |
| Max output | — | 33K |
| Input $ / M tokens | — | $0.04 |
| Output $ / M tokens | — | $2.30 |
| Results tracked | 26 | 31 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.8 27B leads
Qwen2-72B: 29.1 (#310), Qwen3.8 27B: 50.5 (#44)
| Benchmark | Qwen2-72B | Qwen3.8 27B |
|---|---|---|
| LMArena Coding | 1196 | 1482 |
| LMArena WebDev | — | 1593 |
| SciCode | — | 46.6% |
| WeirdML | 11.3% | — |
| BigCodeBench Instruct | 38.5% | — |
| BigCodeBench Complete | 54% | — |
Agentic & Tool Use Qwen3.8 27B leads
Qwen2-72B: 17.0 (#146), Qwen3.8 27B: 32.9 (#57)
| Benchmark | Qwen2-72B | Qwen3.8 27B |
|---|---|---|
| APEX-Agents | — | 47.5% |
| TheAgentCompany | 1.1% | — |
| METR Time Horizons | 29.9% | — |
Reasoning Qwen3.8 27B leads
Qwen2-72B: 23.2 (#181), Qwen3.8 27B: 41.0 (#54)
| Benchmark | Qwen2-72B | Qwen3.8 27B |
|---|---|---|
| LMArena Hard Prompts | 1191 | 1460 |
| Epoch Capabilities Index | 125.28 | 149.38 |
| ARC-AGI-2 | — | 42.4% |
| NYT Connections (extended) | — | 54.5% |
| ARC-AGI-1 | — | 87.5% |
| CritPt | — | 5.4% |
| DTBench | — | 88% |
| LMCA | — | 41.4% |
| Surface Evolver Bench | — | 45% |
Math Qwen3.8 27B leads
Qwen2-72B: 30.2 (#236), Qwen3.8 27B: 37.1 (#161)
| Benchmark | Qwen2-72B | Qwen3.8 27B |
|---|---|---|
| LMArena Math | 1235 | 1456 |
| ProofBench | — | 16% |
| MATH Level 5 | 39.1% | — |
Knowledge Qwen3.8 27B leads
Qwen2-72B: 21.2 (#275), Qwen3.8 27B: 41.6 (#109)
| Benchmark | Qwen2-72B | Qwen3.8 27B |
|---|---|---|
| LMArena Expert | 1171 | 1482 |
| GPQA Diamond | 40.8% | — |
| MMLU | 82.4% | — |
Multimodal Not comparable
Qwen2-72B: —, Qwen3.8 27B: 41.3 (#37)
| Benchmark | Qwen2-72B | Qwen3.8 27B |
|---|---|---|
| LMArena Vision | — | 1271 |
Multilingual Qwen3.8 27B leads
Qwen2-72B: 35.9 (#244), Qwen3.8 27B: 53.7 (#60)
| Benchmark | Qwen2-72B | Qwen3.8 27B |
|---|---|---|
| LMArena Non-English | 1176 | 1430 |
| LMArena Chinese | 1240 | 1504 |
| LMArena French | 1170 | 1465 |
| LMArena German | 1151 | 1438 |
| LMArena Japanese | 1111 | 1384 |
| LMArena Korean | 1083 | 1393 |
| LMArena Russian | 1169 | 1415 |
| LMArena Spanish | 1169 | 1448 |
Instruction Following Qwen3.8 27B leads
Qwen2-72B: 61.7 (#241), Qwen3.8 27B: 75.8 (#53)
| Benchmark | Qwen2-72B | Qwen3.8 27B |
|---|---|---|
| LMArena Instruction Following | 1181 | 1439 |
Long Context Qwen3.8 27B leads
Qwen2-72B: 36.1 (#235), Qwen3.8 27B: 44.3 (#70)
| Benchmark | Qwen2-72B | Qwen3.8 27B |
|---|---|---|
| LMArena Longer Query | 1192 | 1450 |
Writing & Preference Qwen3.8 27B leads
Qwen2-72B: 40.8 (#241), Qwen3.8 27B: 65.8 (#43)
| Benchmark | Qwen2-72B | Qwen3.8 27B |
|---|---|---|
| LMArena Text | 1203 | 1441 |
| LMArena Creative Writing | 1181 | 1384 |
| LMArena Multi-Turn | 1196 | 1441 |
| EQ-Bench Creative Writing | — | 1671 |
Frequently asked questions
Is Qwen2-72B better than Qwen3.8 27B?
Qwen3.8 27B is the stronger model overall, scoring 46.0 to 30.0 on the Noometry Index.
Is Qwen2-72B or Qwen3.8 27B better for coding?
Qwen3.8 27B scores higher on coding benchmarks: 50.5 versus 29.1 in the Noometry coding category.
How many benchmarks do Qwen2-72B and Qwen3.8 27B share?
18 benchmarks have published results for both models. Qwen2-72B has 26 scored results on Noometry and Qwen3.8 27B has 31.