# DeepSeek-V3.1 vs Qwen3 32B

> DeepSeek-V3.1 is the stronger model overall, scoring 42.2 to 38.8 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v3-1-vs-qwen3-32b
- Last updated: 2026-10-10
- Shared benchmarks: 19

## Summary

- They share 19 benchmarks with published results for both. DeepSeek-V3.1 scores higher in 6 categories and Qwen3 32B in 2 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek-V3.1 leads 27.3 to 19.6.
- The biggest single-benchmark swing is Fiction.LiveBench: 52.8% for DeepSeek-V3.1 and 74.2% for Qwen3 32B.
- DeepSeek-V3.1 is cheaper at $0.25 / $0.95 per million input/output tokens, against $0.70 / $2.80 for Qwen3 32B.
- DeepSeek-V3.1 accepts more context: 164K tokens versus 131K.

## Snapshot

| | DeepSeek-V3.1 | Qwen3 32B |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 42.2 | 38.8 |
| Rank | 106 | 166 |
| Context | 164K | 131K |
| Input $/M | $0.25 | $0.70 |
| Output $/M | $0.95 | $2.80 |
| Weights | Open | Open |

## Coding

- DeepSeek-V3.1: 39.8 (#140)
- Qwen3 32B: 37.3 (#184)

| Benchmark | DeepSeek-V3.1 | Qwen3 32B |
|---|---|---|
| LMArena Coding | 1417 | 1358 |
| Aider Polyglot | — | 40% |
| SciCode | — | 35.4% |
| WeirdML | 38.4% | — |

## Agentic & Tool Use

- DeepSeek-V3.1: —
- Qwen3 32B: 32.3 (#58)

| Benchmark | DeepSeek-V3.1 | Qwen3 32B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 48.7% |

## Reasoning

- DeepSeek-V3.1: 27.3 (#109)
- Qwen3 32B: 19.6 (#233)

| Benchmark | DeepSeek-V3.1 | Qwen3 32B |
|---|---|---|
| Kagi LLM Benchmark | 53.2% | 54.9% |
| LMArena Hard Prompts | 1417 | 1335 |
| DTBench | 82.7% | 67.5% |
| LMCA | 24.3% | 17.3% |
| Epoch Capabilities Index | 139.92 | 138.51 |
| SimpleBench | 40% | — |
| CritPt | — | 0.3% |
| Chess Puzzles | — | 5% |
| ForecastBench | 58 | — |

## Math

- DeepSeek-V3.1: 36.1 (#125)
- Qwen3 32B: 37.9 (#89)

| Benchmark | DeepSeek-V3.1 | Qwen3 32B |
|---|---|---|
| LMArena Math | 1420 | 1399 |
| OTIS Mock AIME 2024-2025 | — | 66.9% |

## Knowledge

- DeepSeek-V3.1: 43.0 (#86)
- Qwen3 32B: 40.0 (#112)

| Benchmark | DeepSeek-V3.1 | Qwen3 32B |
|---|---|---|
| Vectara Hallucination Rate | 5.5% | 5.9% |
| LMArena Expert | 1403 | 1362 |
| GPQA Diamond | — | 65.7% |

## Multilingual

- DeepSeek-V3.1: 51.8 (#100)
- Qwen3 32B: 45.8 (#162)

| Benchmark | DeepSeek-V3.1 | Qwen3 32B |
|---|---|---|
| LMArena Non-English | 1400 | 1317 |
| LMArena Chinese | 1470 | 1357 |
| LMArena German | 1412 | 1341 |
| LMArena Russian | 1405 | 1311 |
| LMArena French | 1447 | — |
| LMArena Japanese | 1374 | — |
| LMArena Korean | 1337 | — |
| LMArena Spanish | 1429 | — |

## Instruction Following

- DeepSeek-V3.1: 74.1 (#105)
- Qwen3 32B: 69.1 (#174)

| Benchmark | DeepSeek-V3.1 | Qwen3 32B |
|---|---|---|
| LMArena Instruction Following | 1400 | 1305 |

## Long Context

- DeepSeek-V3.1: 36.3 (#227)
- Qwen3 32B: 43.7 (#83)

| Benchmark | DeepSeek-V3.1 | Qwen3 32B |
|---|---|---|
| Fiction.LiveBench | 52.8% | 74.2% |
| LMArena Longer Query | 1422 | 1327 |

## Writing & Preference

- DeepSeek-V3.1: 60.5 (#93)
- Qwen3 32B: 53.1 (#158)

| Benchmark | DeepSeek-V3.1 | Qwen3 32B |
|---|---|---|
| LMArena Text | 1420 | 1340 |
| LMArena Creative Writing | 1401 | 1297 |
| LMArena Multi-Turn | 1408 | 1331 |
| EQ-Bench Creative Writing | 1436 | — |

## FAQ

### Is DeepSeek-V3.1 better than Qwen3 32B?

DeepSeek-V3.1 is the stronger model overall, scoring 42.2 to 38.8 on the Noometry Index.

### Which is cheaper, DeepSeek-V3.1 or Qwen3 32B?

DeepSeek-V3.1 is cheaper. It lists at $0.25 per million input tokens and $0.95 per million output tokens; Qwen3 32B lists at $0.70 and $2.80.

### Is DeepSeek-V3.1 or Qwen3 32B better for coding?

DeepSeek-V3.1 scores higher on coding benchmarks: 39.8 versus 37.3 in the Noometry coding category.

### Which has the bigger context window?

DeepSeek-V3.1 does, with 164K tokens against 131K.

### How many benchmarks do DeepSeek-V3.1 and Qwen3 32B share?

19 benchmarks have published results for both models. DeepSeek-V3.1 has 27 scored results on Noometry and Qwen3 32B has 26.