# Qwen2.5-Coder-32B vs Qwen3-30B-A3B

> Qwen3-30B-A3B is the stronger model overall, scoring 38.9 to 33.4 on the Noometry Index.

- Canonical page: https://noometry.com/compare/qwen2-5-coder-32b-vs-qwen3-30b-a3b
- Last updated: 2026-10-12
- Shared benchmarks: 13

## Summary

- They share 13 benchmarks with published results for both. Qwen2.5-Coder-32B scores higher in 1 category and Qwen3-30B-A3B in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in coding, where Qwen3-30B-A3B leads 37.5 to 22.6.
- Qwen3-30B-A3B is cheaper at $0.12 / $0.50 per million input/output tokens, against $0.66 / $1 for Qwen2.5-Coder-32B.
- Qwen3-30B-A3B accepts more context: 41K tokens versus 33K.

## Snapshot

| | Qwen2.5-Coder-32B | Qwen3-30B-A3B |
|---|---|---|
| Provider | Alibaba (Qwen) | Alibaba (Qwen) |
| Noometry Index | 33.4 | 38.9 |
| Rank | 245 | 179 |
| Context | 33K | 41K |
| Input $/M | $0.66 | $0.12 |
| Output $/M | $1 | $0.50 |
| Weights | Open | Open |

## Coding

- Qwen2.5-Coder-32B: 22.6 (#333)
- Qwen3-30B-A3B: 37.5 (#194)

| Benchmark | Qwen2.5-Coder-32B | Qwen3-30B-A3B |
|---|---|---|
| LMArena Coding | 1276 | 1416 |
| SWE-bench Verified (bash only) | 9% | — |
| Aider Polyglot | 16.4% | — |
| SciCode | — | 33.3% |
| WeirdML | — | 29.8% |
| BigCodeBench Instruct | 49% | — |
| LiveBench Coding | 56.9% | — |
| BigCodeBench Complete | 58% | — |
| HumanEval+ | 87.2% | — |
| MBPP+ | 77% | — |

## Agentic & Tool Use

- Qwen2.5-Coder-32B: —
- Qwen3-30B-A3B: 29.8 (#82)

| Benchmark | Qwen2.5-Coder-32B | Qwen3-30B-A3B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 41.4% |

## Reasoning

- Qwen2.5-Coder-32B: 21.2 (#225)
- Qwen3-30B-A3B: 22.2 (#204)

| Benchmark | Qwen2.5-Coder-32B | Qwen3-30B-A3B |
|---|---|---|
| LMArena Hard Prompts | 1251 | 1398 |
| Epoch Capabilities Index | 119.49 | 139.63 |
| Kagi LLM Benchmark | — | 54.9% |
| CritPt | — | 0.3% |
| Chess Puzzles | — | 8% |
| LiveBench Reasoning | 42.1% | — |
| DTBench | — | 69.3% |
| LiveBench Data Analysis | 49.9% | — |
| LMCA | — | 22.4% |
| HellaSwag | 83% | — |
| LiveBench | 46.2% | — |
| WinoGrande | 80.8% | — |

## Math

- Qwen2.5-Coder-32B: 33.3 (#204)
- Qwen3-30B-A3B: 37.4 (#157)

| Benchmark | Qwen2.5-Coder-32B | Qwen3-30B-A3B |
|---|---|---|
| LMArena Math | 1251 | 1394 |
| MathArena Final-Answer Competitions | — | 47.8% |
| OTIS Mock AIME 2024-2025 | — | 70.3% |
| LiveBench Math | 46.6% | — |
| GSM8K | 93% | — |

## Knowledge

- Qwen2.5-Coder-32B: 33.4 (#203)
- Qwen3-30B-A3B: 41.8 (#105)

| Benchmark | Qwen2.5-Coder-32B | Qwen3-30B-A3B |
|---|---|---|
| LMArena Expert | 1221 | 1396 |
| GPQA Diamond | — | 70.1% |
| Confabulations | — | 12.3% |
| ARC (AI2) Challenge | 70.5% | — |
| MMLU | 79.1% | — |

## Multilingual

- Qwen2.5-Coder-32B: 37.8 (#235)
- Qwen3-30B-A3B: 49.5 (#132)

| Benchmark | Qwen2.5-Coder-32B | Qwen3-30B-A3B |
|---|---|---|
| LMArena Non-English | 1205 | 1372 |
| LMArena Chinese | 1222 | 1433 |
| LMArena Russian | 1228 | 1370 |
| LMArena French | — | 1418 |
| LMArena German | — | 1380 |
| LMArena Japanese | — | 1337 |
| LMArena Korean | — | 1331 |
| LMArena Spanish | — | 1404 |

## Instruction Following

- Qwen2.5-Coder-32B: 61.4 (#245)
- Qwen3-30B-A3B: 72.0 (#142)

| Benchmark | Qwen2.5-Coder-32B | Qwen3-30B-A3B |
|---|---|---|
| LMArena Instruction Following | 1223 | 1363 |
| LiveBench Instruction Following | 58.7% | — |

## Long Context

- Qwen2.5-Coder-32B: 38.0 (#208)
- Qwen3-30B-A3B: 31.0 (#283)

| Benchmark | Qwen2.5-Coder-32B | Qwen3-30B-A3B |
|---|---|---|
| LMArena Longer Query | 1251 | 1379 |
| Fiction.LiveBench | — | 40.6% |

## Writing & Preference

- Qwen2.5-Coder-32B: 41.6 (#240)
- Qwen3-30B-A3B: 55.6 (#143)

| Benchmark | Qwen2.5-Coder-32B | Qwen3-30B-A3B |
|---|---|---|
| LMArena Text | 1230 | 1384 |
| LMArena Creative Writing | 1174 | 1317 |
| LMArena Multi-Turn | 1222 | 1378 |
| Short-Story Creative Writing | — | 75.3% |
| LiveBench Language | 23.3% | — |

## FAQ

### Is Qwen2.5-Coder-32B better than Qwen3-30B-A3B?

Qwen3-30B-A3B is the stronger model overall, scoring 38.9 to 33.4 on the Noometry Index.

### Which is cheaper, Qwen2.5-Coder-32B or Qwen3-30B-A3B?

Qwen3-30B-A3B is cheaper. It lists at $0.12 per million input tokens and $0.50 per million output tokens; Qwen2.5-Coder-32B lists at $0.66 and $1.

### Is Qwen2.5-Coder-32B or Qwen3-30B-A3B better for coding?

Qwen3-30B-A3B scores higher on coding benchmarks: 37.5 versus 22.6 in the Noometry coding category.

### Which has the bigger context window?

Qwen3-30B-A3B does, with 41K tokens against 33K.

### How many benchmarks do Qwen2.5-Coder-32B and Qwen3-30B-A3B share?

13 benchmarks have published results for both models. Qwen2.5-Coder-32B has 31 scored results on Noometry and Qwen3-30B-A3B has 32.