Model comparison
Qwen2.5-Coder-32B vs Qwen3 235B-A22B
Qwen3 235B-A22B is the stronger model overall, scoring 43.5 to 33.4 on the Noometry Index. Qwen2.5-Coder-32B costs 1.6× less per token, which makes it the better buy when Qwen3 235B-A22B's lead doesn't matter for your workload.
Last verified . 14 shared benchmarks.
Alibaba (Qwen)Summary
- They share 14 benchmarks with published results for both. Qwen2.5-Coder-32B scores higher in 1 category and Qwen3 235B-A22B in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in coding, where Qwen3 235B-A22B leads 44.3 to 22.6.
- The biggest single-benchmark swing is Aider Polyglot: 16.4% for Qwen2.5-Coder-32B and 59.6% for Qwen3 235B-A22B.
- Qwen2.5-Coder-32B is cheaper at $0.66 / $1 per million input/output tokens, against $0.70 / $2.80 for Qwen3 235B-A22B.
- Qwen3 235B-A22B accepts more context: 131K tokens versus 33K.
Side by side
| Qwen2.5-Coder-32B | Qwen3 235B-A22B | |
|---|---|---|
| Provider | Alibaba (Qwen) | Alibaba (Qwen) |
| Noometry Index | 33.4 | 43.5 |
| Released | 2024-09-18 | 2025-04 |
| Weights | Open | Open |
| Context window | 33K | 131K |
| Max output | 29K | 16K |
| Input $ / M tokens | $0.66 | $0.70 |
| Output $ / M tokens | $1 | $2.80 |
| Results tracked | 31 | 49 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3 235B-A22B leads
Qwen2.5-Coder-32B: 22.6 (#333), Qwen3 235B-A22B: 44.3 (#75)
| Benchmark | Qwen2.5-Coder-32B | Qwen3 235B-A22B |
|---|---|---|
| Aider Polyglot | 16.4% | 59.6% |
| LMArena Coding | 1276 | 1445 |
| SWE-bench Verified (bash only) | 9% | — |
| SciCode | — | 42.4% |
| WeirdML | — | 41% |
| 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 235B-A22B: 33.9 (#51)
| Benchmark | Qwen2.5-Coder-32B | Qwen3 235B-A22B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 52.1% |
| Vending-Bench 2 | — | -11.34 |
Reasoning Qwen2.5-Coder-32B leads
Qwen2.5-Coder-32B: 21.2 (#225), Qwen3 235B-A22B: 15.7 (#311)
| Benchmark | Qwen2.5-Coder-32B | Qwen3 235B-A22B |
|---|---|---|
| LMArena Hard Prompts | 1251 | 1433 |
| Epoch Capabilities Index | 119.49 | 143.85 |
| ARC-AGI-2 | — | 1.3% |
| SimpleBench | — | 31% |
| Kagi LLM Benchmark | — | 69.4% |
| ARC-AGI-1 | — | 11% |
| CritPt | — | 0% |
| Chess Puzzles | — | 12% |
| LiveBench Reasoning | 42.1% | — |
| Mystery Game Puzzles | — | 9% |
| DTBench | — | 80.3% |
| LiveBench Data Analysis | 49.9% | — |
| LMCA | — | 29.3% |
| ForecastBench | — | 59.7 |
| HellaSwag | 83% | — |
| LiveBench | 46.2% | — |
| WinoGrande | 80.8% | — |
Math Qwen3 235B-A22B leads
Qwen2.5-Coder-32B: 33.3 (#204), Qwen3 235B-A22B: 50.4 (#57)
| Benchmark | Qwen2.5-Coder-32B | Qwen3 235B-A22B |
|---|---|---|
| LMArena Math | 1251 | 1432 |
| OTIS Mock AIME 2024-2025 | — | 86.7% |
| Omni-MATH | — | 71.8% |
| LiveBench Math | 46.6% | — |
| MATH Level 5 | — | 68.9% |
| FrontierMath (Feb 2025 set) | — | 8.5% |
| FrontierMath Tier 4 (v1) | — | 0% |
| GSM8K | 93% | — |
Knowledge Qwen3 235B-A22B leads
Qwen2.5-Coder-32B: 33.4 (#203), Qwen3 235B-A22B: 49.6 (#73)
| Benchmark | Qwen2.5-Coder-32B | Qwen3 235B-A22B |
|---|---|---|
| LMArena Expert | 1221 | 1463 |
| GPQA Diamond | — | 80.1% |
| SimpleQA Verified | — | 40.4% |
| MMLU-Pro | — | 84.4% |
| Confabulations | — | 15.6% |
| Vectara Hallucination Rate | — | 9.3% |
| GPQA (HELM) | — | 72.7% |
| ARC (AI2) Challenge | 70.5% | — |
| MMLU | 79.1% | — |
Multilingual Qwen3 235B-A22B leads
Qwen2.5-Coder-32B: 37.8 (#235), Qwen3 235B-A22B: 52.3 (#89)
| Benchmark | Qwen2.5-Coder-32B | Qwen3 235B-A22B |
|---|---|---|
| LMArena Non-English | 1205 | 1409 |
| LMArena Chinese | 1222 | 1481 |
| LMArena Russian | 1228 | 1411 |
| LMArena French | — | 1445 |
| LMArena German | — | 1433 |
| LMArena Japanese | — | 1399 |
| LMArena Korean | — | 1391 |
| LMArena Spanish | — | 1430 |
Instruction Following Qwen3 235B-A22B leads
Qwen2.5-Coder-32B: 61.4 (#245), Qwen3 235B-A22B: 72.6 (#136)
| Benchmark | Qwen2.5-Coder-32B | Qwen3 235B-A22B |
|---|---|---|
| LMArena Instruction Following | 1223 | 1408 |
| LiveBench Instruction Following | 58.7% | — |
| IFEval | — | 83.5% |
Long Context Qwen3 235B-A22B leads
Qwen2.5-Coder-32B: 38.0 (#208), Qwen3 235B-A22B: 46.1 (#26)
| Benchmark | Qwen2.5-Coder-32B | Qwen3 235B-A22B |
|---|---|---|
| LMArena Longer Query | 1251 | 1426 |
| Fiction.LiveBench | — | 75% |
Writing & Preference Qwen3 235B-A22B leads
Qwen2.5-Coder-32B: 41.6 (#240), Qwen3 235B-A22B: 59.6 (#108)
| Benchmark | Qwen2.5-Coder-32B | Qwen3 235B-A22B |
|---|---|---|
| LMArena Text | 1230 | 1419 |
| LMArena Creative Writing | 1174 | 1384 |
| LMArena Multi-Turn | 1222 | 1432 |
| Short-Story Creative Writing | — | 83% |
| EQ-Bench Creative Writing | — | 1366 |
| WildBench | — | 86.6% |
| LiveBench Language | 23.3% | — |
Frequently asked questions
Is Qwen2.5-Coder-32B better than Qwen3 235B-A22B?
Qwen3 235B-A22B is the stronger model overall, scoring 43.5 to 33.4 on the Noometry Index. Qwen2.5-Coder-32B costs 1.6× less per token, which makes it the better buy when Qwen3 235B-A22B's lead doesn't matter for your workload.
Which is cheaper, Qwen2.5-Coder-32B or Qwen3 235B-A22B?
Qwen2.5-Coder-32B is cheaper. It lists at $0.66 per million input tokens and $1 per million output tokens; Qwen3 235B-A22B lists at $0.70 and $2.80.
Is Qwen2.5-Coder-32B or Qwen3 235B-A22B better for coding?
Qwen3 235B-A22B scores higher on coding benchmarks: 44.3 versus 22.6 in the Noometry coding category.
Which has the bigger context window?
Qwen3 235B-A22B does, with 131K tokens against 33K.
How many benchmarks do Qwen2.5-Coder-32B and Qwen3 235B-A22B share?
14 benchmarks have published results for both models. Qwen2.5-Coder-32B has 31 scored results on Noometry and Qwen3 235B-A22B has 49.