Model comparison
Qwen2.5-Coder-32B vs Qwen3.8 27B
Qwen3.8 27B is the stronger model overall, scoring 46.0 to 33.4 on the Noometry Index.
Last verified . 13 shared benchmarks.
Alibaba (Qwen)Summary
- They share 13 benchmarks with published results for both. Qwen2.5-Coder-32B scores higher in 0 categories and Qwen3.8 27B in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in coding, where Qwen3.8 27B leads 50.5 to 22.6.
- Qwen2.5-Coder-32B is cheaper at $0.66 / $1 per million input/output tokens, against $0.99 / $1.49 for Qwen3.8 27B.
- Qwen3.8 27B accepts more context: 262K tokens versus 33K.
Side by side
| Qwen2.5-Coder-32B | Qwen3.8 27B | |
|---|---|---|
| Provider | Alibaba (Qwen) | Alibaba (Qwen) |
| Noometry Index | 33.4 | 46.0 |
| Released | 2024-09-18 | 2026-08-14 |
| Weights | Open | Open |
| Context window | 33K | 262K |
| Max output | 29K | 33K |
| Input $ / M tokens | $0.66 | $0.99 |
| Output $ / M tokens | $1 | $1.49 |
| Results tracked | 31 | 31 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.8 27B leads
Qwen2.5-Coder-32B: 22.6 (#333), Qwen3.8 27B: 50.5 (#44)
| Benchmark | Qwen2.5-Coder-32B | Qwen3.8 27B |
|---|---|---|
| LMArena Coding | 1276 | 1482 |
| SWE-bench Verified (bash only) | 9% | — |
| Aider Polyglot | 16.4% | — |
| LMArena WebDev | — | 1593 |
| SciCode | — | 46.6% |
| BigCodeBench Instruct | 49% | — |
| LiveBench Coding | 56.9% | — |
| BigCodeBench Complete | 58% | — |
| HumanEval+ | 87.2% | — |
| MBPP+ | 77% | — |
Agentic & Tool Use Not comparable
Qwen2.5-Coder-32B: —, Qwen3.8 27B: 32.9 (#57)
| Benchmark | Qwen2.5-Coder-32B | Qwen3.8 27B |
|---|---|---|
| APEX-Agents | — | 47.5% |
Reasoning Qwen3.8 27B leads
Qwen2.5-Coder-32B: 21.2 (#225), Qwen3.8 27B: 41.0 (#54)
| Benchmark | Qwen2.5-Coder-32B | Qwen3.8 27B |
|---|---|---|
| LMArena Hard Prompts | 1251 | 1460 |
| Epoch Capabilities Index | 119.49 | 149.38 |
| ARC-AGI-2 | — | 42.4% |
| NYT Connections (extended) | — | 54.5% |
| ARC-AGI-1 | — | 87.5% |
| CritPt | — | 5.4% |
| LiveBench Reasoning | 42.1% | — |
| DTBench | — | 88% |
| LiveBench Data Analysis | 49.9% | — |
| LMCA | — | 41.4% |
| Surface Evolver Bench | — | 45% |
| HellaSwag | 83% | — |
| LiveBench | 46.2% | — |
| WinoGrande | 80.8% | — |
Math Qwen3.8 27B leads
Qwen2.5-Coder-32B: 33.3 (#204), Qwen3.8 27B: 37.1 (#161)
| Benchmark | Qwen2.5-Coder-32B | Qwen3.8 27B |
|---|---|---|
| LMArena Math | 1251 | 1456 |
| ProofBench | — | 16% |
| LiveBench Math | 46.6% | — |
| GSM8K | 93% | — |
Knowledge Qwen3.8 27B leads
Qwen2.5-Coder-32B: 33.4 (#203), Qwen3.8 27B: 41.6 (#109)
| Benchmark | Qwen2.5-Coder-32B | Qwen3.8 27B |
|---|---|---|
| LMArena Expert | 1221 | 1482 |
| ARC (AI2) Challenge | 70.5% | — |
| MMLU | 79.1% | — |
Multimodal Not comparable
Qwen2.5-Coder-32B: —, Qwen3.8 27B: 41.3 (#37)
| Benchmark | Qwen2.5-Coder-32B | Qwen3.8 27B |
|---|---|---|
| LMArena Vision | — | 1271 |
Multilingual Qwen3.8 27B leads
Qwen2.5-Coder-32B: 37.8 (#235), Qwen3.8 27B: 53.7 (#60)
| Benchmark | Qwen2.5-Coder-32B | Qwen3.8 27B |
|---|---|---|
| LMArena Non-English | 1205 | 1430 |
| LMArena Chinese | 1222 | 1504 |
| LMArena Russian | 1228 | 1415 |
| LMArena French | — | 1465 |
| LMArena German | — | 1438 |
| LMArena Japanese | — | 1384 |
| LMArena Korean | — | 1393 |
| LMArena Spanish | — | 1448 |
Instruction Following Qwen3.8 27B leads
Qwen2.5-Coder-32B: 61.4 (#245), Qwen3.8 27B: 75.8 (#53)
| Benchmark | Qwen2.5-Coder-32B | Qwen3.8 27B |
|---|---|---|
| LMArena Instruction Following | 1223 | 1439 |
| LiveBench Instruction Following | 58.7% | — |
Long Context Qwen3.8 27B leads
Qwen2.5-Coder-32B: 38.0 (#208), Qwen3.8 27B: 44.3 (#70)
| Benchmark | Qwen2.5-Coder-32B | Qwen3.8 27B |
|---|---|---|
| LMArena Longer Query | 1251 | 1450 |
Writing & Preference Qwen3.8 27B leads
Qwen2.5-Coder-32B: 41.6 (#240), Qwen3.8 27B: 65.8 (#43)
| Benchmark | Qwen2.5-Coder-32B | Qwen3.8 27B |
|---|---|---|
| LMArena Text | 1230 | 1441 |
| LMArena Creative Writing | 1174 | 1384 |
| LMArena Multi-Turn | 1222 | 1441 |
| EQ-Bench Creative Writing | — | 1671 |
| LiveBench Language | 23.3% | — |
Frequently asked questions
Is Qwen2.5-Coder-32B better than Qwen3.8 27B?
Qwen3.8 27B is the stronger model overall, scoring 46.0 to 33.4 on the Noometry Index.
Which is cheaper, Qwen2.5-Coder-32B or Qwen3.8 27B?
Qwen2.5-Coder-32B is cheaper. It lists at $0.66 per million input tokens and $1 per million output tokens; Qwen3.8 27B lists at $0.99 and $1.49.
Is Qwen2.5-Coder-32B or Qwen3.8 27B better for coding?
Qwen3.8 27B scores higher on coding benchmarks: 50.5 versus 22.6 in the Noometry coding category.
Which has the bigger context window?
Qwen3.8 27B does, with 262K tokens against 33K.
How many benchmarks do Qwen2.5-Coder-32B and Qwen3.8 27B share?
13 benchmarks have published results for both models. Qwen2.5-Coder-32B has 31 scored results on Noometry and Qwen3.8 27B has 31.