# DeepSeek-V2.5 (Sep 2024) vs Qwen3 32B

> Qwen3 32B is the stronger model overall, scoring 38.8 to 37.0 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v2-5-vs-qwen3-32b
- Last updated: 2026-10-10
- Shared benchmarks: 14

## Summary

- They share 14 benchmarks with published results for both. DeepSeek-V2.5 (Sep 2024) scores higher in 1 category and Qwen3 32B in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in coding, where Qwen3 32B leads 37.3 to 31.3.
- The biggest single-benchmark swing is Aider Polyglot: 17.8% for DeepSeek-V2.5 (Sep 2024) and 40% for Qwen3 32B.

## Snapshot

| | DeepSeek-V2.5 (Sep 2024) | Qwen3 32B |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 37.0 | 38.8 |
| Rank | 197 | 166 |
| Context | — | 131K |
| Input $/M | — | $0.70 |
| Output $/M | — | $2.80 |
| Weights | Open | Open |

## Coding

- DeepSeek-V2.5 (Sep 2024): 31.3 (#274)
- Qwen3 32B: 37.3 (#184)

| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen3 32B |
|---|---|---|
| Aider Polyglot | 17.8% | 40% |
| LMArena Coding | 1310 | 1358 |
| SciCode | — | 35.4% |
| BigCodeBench Instruct | 48.6% | — |
| BigCodeBench Complete | 53.2% | — |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |

## Agentic & Tool Use

- DeepSeek-V2.5 (Sep 2024): —
- Qwen3 32B: 32.3 (#58)

| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen3 32B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 48.7% |

## Reasoning

- DeepSeek-V2.5 (Sep 2024): 24.5 (#145)
- Qwen3 32B: 19.6 (#233)

| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen3 32B |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1335 |
| Kagi LLM Benchmark | — | 54.9% |
| CritPt | — | 0.3% |
| Chess Puzzles | — | 5% |
| DTBench | — | 67.5% |
| LMCA | — | 17.3% |
| Epoch Capabilities Index | — | 138.51 |

## Math

- DeepSeek-V2.5 (Sep 2024): 33.3 (#177)
- Qwen3 32B: 37.9 (#89)

| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen3 32B |
|---|---|---|
| LMArena Math | 1289 | 1399 |
| OTIS Mock AIME 2024-2025 | — | 66.9% |

## Knowledge

- DeepSeek-V2.5 (Sep 2024): 34.2 (#190)
- Qwen3 32B: 40.0 (#112)

| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen3 32B |
|---|---|---|
| LMArena Expert | 1266 | 1362 |
| GPQA Diamond | — | 65.7% |
| Vectara Hallucination Rate | — | 5.9% |

## Multilingual

- DeepSeek-V2.5 (Sep 2024): 42.7 (#188)
- Qwen3 32B: 45.8 (#162)

| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen3 32B |
|---|---|---|
| LMArena Non-English | 1273 | 1317 |
| LMArena Chinese | 1317 | 1357 |
| LMArena German | 1258 | 1341 |
| LMArena Russian | 1289 | 1311 |
| LMArena French | 1288 | — |
| LMArena Japanese | 1228 | — |
| LMArena Korean | 1209 | — |
| LMArena Spanish | 1248 | — |

## Instruction Following

- DeepSeek-V2.5 (Sep 2024): 67.7 (#189)
- Qwen3 32B: 69.1 (#174)

| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen3 32B |
|---|---|---|
| LMArena Instruction Following | 1280 | 1305 |

## Long Context

- DeepSeek-V2.5 (Sep 2024): 39.5 (#169)
- Qwen3 32B: 43.7 (#83)

| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen3 32B |
|---|---|---|
| LMArena Longer Query | 1301 | 1327 |
| Fiction.LiveBench | — | 74.2% |

## Writing & Preference

- DeepSeek-V2.5 (Sep 2024): 50.0 (#182)
- Qwen3 32B: 53.1 (#158)

| Benchmark | DeepSeek-V2.5 (Sep 2024) | Qwen3 32B |
|---|---|---|
| LMArena Text | 1294 | 1340 |
| LMArena Creative Writing | 1285 | 1297 |
| LMArena Multi-Turn | 1297 | 1331 |

## FAQ

### Is DeepSeek-V2.5 (Sep 2024) better than Qwen3 32B?

Qwen3 32B is the stronger model overall, scoring 38.8 to 37.0 on the Noometry Index.

### Is DeepSeek-V2.5 (Sep 2024) or Qwen3 32B better for coding?

Qwen3 32B scores higher on coding benchmarks: 37.3 versus 31.3 in the Noometry coding category.

### How many benchmarks do DeepSeek-V2.5 (Sep 2024) and Qwen3 32B share?

14 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and Qwen3 32B has 26.