# DeepSeek V4 Pro vs Qwen3 32B

> DeepSeek V4 Pro is the stronger model overall, scoring 54.2 to 38.8 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v4-pro-vs-qwen3-32b
- Last updated: 2026-10-10
- Shared benchmarks: 23

## Summary

- They share 23 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 9 categories and Qwen3 32B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Pro leads 56.4 to 19.6.
- The biggest single-benchmark swing is Chess Puzzles: 47% for DeepSeek V4 Pro and 5% for Qwen3 32B.
- DeepSeek V4 Pro is cheaper at $0.43 / $0.87 per million input/output tokens, against $0.70 / $2.80 for Qwen3 32B.
- DeepSeek V4 Pro accepts more context: 1M tokens versus 131K.

## Snapshot

| | DeepSeek V4 Pro | Qwen3 32B |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 54.2 | 38.8 |
| Rank | 28 | 166 |
| Context | 1M | 131K |
| Input $/M | $0.43 | $0.70 |
| Output $/M | $0.87 | $2.80 |
| Weights | Open | Open |

## Coding

- DeepSeek V4 Pro: 52.2 (#31)
- Qwen3 32B: 37.3 (#184)

| Benchmark | DeepSeek V4 Pro | Qwen3 32B |
|---|---|---|
| SciCode | 51% | 35.4% |
| LMArena Coding | 1470 | 1358 |
| SWE-bench Verified | 77.6% | — |
| FrontierCode | 28.6% | — |
| Aider Polyglot | — | 40% |
| LMArena WebDev | 1582 | — |
| WeirdML | 66.2% | — |
| ALE-Bench | 1,403 | — |

## Agentic & Tool Use

- DeepSeek V4 Pro: 32.7 (#54)
- Qwen3 32B: 32.3 (#58)

| Benchmark | DeepSeek V4 Pro | Qwen3 32B |
|---|---|---|
| APEX-Agents | 47.3% | — |
| Berkeley Function Calling Leaderboard | — | 48.7% |
| Vending-Bench 2 | 3,285 | — |

## Reasoning

- DeepSeek V4 Pro: 56.4 (#22)
- Qwen3 32B: 19.6 (#233)

| Benchmark | DeepSeek V4 Pro | Qwen3 32B |
|---|---|---|
| Kagi LLM Benchmark | 53.5% | 54.9% |
| CritPt | 18% | 0.3% |
| Chess Puzzles | 47% | 5% |
| LMArena Hard Prompts | 1461 | 1335 |
| DTBench | 93.9% | 67.5% |
| LMCA | 45.5% | 17.3% |
| Epoch Capabilities Index | 155.39 | 138.51 |
| ARC-AGI-2 | 61.3% | — |
| NYT Connections (extended) | 91.3% | — |
| ARC-AGI-1 | 90.5% | — |
| Mystery Game Puzzles | 43% | — |
| Surface Evolver Bench | 40% | — |
| ForecastBench | 56.1 | — |

## Math

- DeepSeek V4 Pro: 64.6 (#24)
- Qwen3 32B: 37.9 (#89)

| Benchmark | DeepSeek V4 Pro | Qwen3 32B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 98.6% | 66.9% |
| LMArena Math | 1455 | 1399 |
| FrontierMath (Tiers 1-3) | 64.6% | — |
| FrontierMath Tier 4 | 26.8% | — |
| MathArena Final-Answer Competitions | 76.6% | — |
| ProofBench | 50% | — |

## Knowledge

- DeepSeek V4 Pro: 59.5 (#28)
- Qwen3 32B: 40.0 (#112)

| Benchmark | DeepSeek V4 Pro | Qwen3 32B |
|---|---|---|
| GPQA Diamond | 91.7% | 65.7% |
| Vectara Hallucination Rate | 8.6% | 5.9% |
| LMArena Expert | 1463 | 1362 |
| SimpleQA Verified | 52.9% | — |

## Multilingual

- DeepSeek V4 Pro: 54.5 (#44)
- Qwen3 32B: 45.8 (#162)

| Benchmark | DeepSeek V4 Pro | Qwen3 32B |
|---|---|---|
| LMArena Non-English | 1438 | 1317 |
| LMArena Chinese | 1486 | 1357 |
| LMArena German | 1458 | 1341 |
| LMArena Russian | 1452 | 1311 |
| LMArena French | 1473 | — |
| LMArena Japanese | 1448 | — |
| LMArena Korean | 1452 | — |
| LMArena Spanish | 1457 | — |

## Instruction Following

- DeepSeek V4 Pro: 76.3 (#44)
- Qwen3 32B: 69.1 (#174)

| Benchmark | DeepSeek V4 Pro | Qwen3 32B |
|---|---|---|
| LMArena Instruction Following | 1448 | 1305 |

## Long Context

- DeepSeek V4 Pro: 45.0 (#49)
- Qwen3 32B: 43.7 (#83)

| Benchmark | DeepSeek V4 Pro | Qwen3 32B |
|---|---|---|
| LMArena Longer Query | 1459 | 1327 |
| Fiction.LiveBench | — | 74.2% |
| CL-bench Life | 13.5% | — |

## Writing & Preference

- DeepSeek V4 Pro: 65.6 (#45)
- Qwen3 32B: 53.1 (#158)

| Benchmark | DeepSeek V4 Pro | Qwen3 32B |
|---|---|---|
| LMArena Text | 1451 | 1340 |
| LMArena Creative Writing | 1446 | 1297 |
| LMArena Multi-Turn | 1467 | 1331 |
| EQ-Bench Creative Writing | 1553 | — |
| EQ-Bench 4 | 1166 | — |

## FAQ

### Is DeepSeek V4 Pro better than Qwen3 32B?

DeepSeek V4 Pro is the stronger model overall, scoring 54.2 to 38.8 on the Noometry Index.

### Which is cheaper, DeepSeek V4 Pro or Qwen3 32B?

DeepSeek V4 Pro is cheaper. It lists at $0.43 per million input tokens and $0.87 per million output tokens; Qwen3 32B lists at $0.70 and $2.80.

### Is DeepSeek V4 Pro or Qwen3 32B better for coding?

DeepSeek V4 Pro scores higher on coding benchmarks: 52.2 versus 37.3 in the Noometry coding category.

### Which has the bigger context window?

DeepSeek V4 Pro does, with 1M tokens against 131K.

### How many benchmarks do DeepSeek V4 Pro and Qwen3 32B share?

23 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and Qwen3 32B has 26.