# DeepSeek V4.1 Flash vs Qwen3-4B

> DeepSeek V4.1 Flash is the stronger model overall, scoring 52.8 to 31.9 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v4-1-flash-vs-qwen3-4b
- Last updated: 2026-10-10
- Shared benchmarks: 2

## Summary

- They share 2 benchmarks with published results for both. DeepSeek V4.1 Flash scores higher in 4 categories and Qwen3-4B in 0 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek V4.1 Flash leads 66.7 to 29.7.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 98.3% for DeepSeek V4.1 Flash and 52.2% for Qwen3-4B.

## Snapshot

| | DeepSeek V4.1 Flash | Qwen3-4B |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 52.8 | 31.9 |
| Rank | 38 | 264 |
| Context | 1M | — |
| Input $/M | $0.15 | — |
| Output $/M | $0.60 | — |
| Weights | Open | Open |

## Coding

- DeepSeek V4.1 Flash: 52.9 (#32)
- Qwen3-4B: —

| Benchmark | DeepSeek V4.1 Flash | Qwen3-4B |
|---|---|---|
| LMArena WebDev | 1619 | — |
| SciCode | 51.9% | — |
| LMArena Coding | 1506 | — |
| ALE-Bench | 1,092 | — |

## Agentic & Tool Use

- DeepSeek V4.1 Flash: 31.2 (#69)
- Qwen3-4B: 27.6 (#100)

| Benchmark | DeepSeek V4.1 Flash | Qwen3-4B |
|---|---|---|
| APEX-Agents | 39.5% | — |
| Berkeley Function Calling Leaderboard | — | 35.7% |
| GDP.pdf | 19.8% | — |

## Reasoning

- DeepSeek V4.1 Flash: 50.2 (#36)
- Qwen3-4B: 19.2 (#268)

| Benchmark | DeepSeek V4.1 Flash | Qwen3-4B |
|---|---|---|
| NYT Connections (extended) | 89.6% | — |
| CritPt | 14.3% | — |
| Chess Puzzles | — | 4% |
| LMArena Hard Prompts | 1483 | — |
| Mystery Game Puzzles | 43% | — |
| DTBench | 89.9% | — |
| LMCA | 47% | — |
| Surface Evolver Bench | 46.3% | — |
| Epoch Capabilities Index | 154.9 | — |

## Math

- DeepSeek V4.1 Flash: 66.7 (#25)
- Qwen3-4B: 29.7 (#240)

| Benchmark | DeepSeek V4.1 Flash | Qwen3-4B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 98.3% | 52.2% |
| FrontierMath (Tiers 1-3) | 67.4% | — |
| FrontierMath Tier 4 | 26.8% | — |
| MathArena Final-Answer Competitions | — | 38.5% |
| ProofBench | 54% | — |
| LMArena Math | 1477 | — |

## Knowledge

- DeepSeek V4.1 Flash: 57.9 (#38)
- Qwen3-4B: 33.0 (#208)

| Benchmark | DeepSeek V4.1 Flash | Qwen3-4B |
|---|---|---|
| GPQA Diamond | 89.8% | 52.3% |
| Vectara Hallucination Rate | — | 5.7% |
| LMArena Expert | 1506 | — |

## Multimodal

- DeepSeek V4.1 Flash: 39.1 (#61)
- Qwen3-4B: —

| Benchmark | DeepSeek V4.1 Flash | Qwen3-4B |
|---|---|---|
| LMArena Vision | 1277 | — |
| Furniture Assembly | 34.2% | — |

## Multilingual

- DeepSeek V4.1 Flash: 55.0 (#35)
- Qwen3-4B: —

| Benchmark | DeepSeek V4.1 Flash | Qwen3-4B |
|---|---|---|
| LMArena Non-English | 1448 | — |
| LMArena Chinese | 1497 | — |
| LMArena French | 1452 | — |
| LMArena German | 1484 | — |
| LMArena Japanese | 1412 | — |
| LMArena Korean | 1452 | — |
| LMArena Russian | 1471 | — |
| LMArena Spanish | 1459 | — |

## Instruction Following

- DeepSeek V4.1 Flash: 77.3 (#26)
- Qwen3-4B: —

| Benchmark | DeepSeek V4.1 Flash | Qwen3-4B |
|---|---|---|
| LMArena Instruction Following | 1474 | — |

## Long Context

- DeepSeek V4.1 Flash: 45.2 (#47)
- Qwen3-4B: —

| Benchmark | DeepSeek V4.1 Flash | Qwen3-4B |
|---|---|---|
| LMArena Longer Query | 1475 | — |

## Writing & Preference

- DeepSeek V4.1 Flash: 65.4 (#48)
- Qwen3-4B: —

| Benchmark | DeepSeek V4.1 Flash | Qwen3-4B |
|---|---|---|
| LMArena Text | 1462 | — |
| LMArena Creative Writing | 1435 | — |
| EQ-Bench Creative Writing | 1540 | — |
| LMArena Multi-Turn | 1457 | — |

## FAQ

### Is DeepSeek V4.1 Flash better than Qwen3-4B?

DeepSeek V4.1 Flash is the stronger model overall, scoring 52.8 to 31.9 on the Noometry Index.

### How many benchmarks do DeepSeek V4.1 Flash and Qwen3-4B share?

2 benchmarks have published results for both models. DeepSeek V4.1 Flash has 37 scored results on Noometry and Qwen3-4B has 6.