# DeepSeek-V3 vs Qwen2.5 14B Instruct

> DeepSeek-V3 has enough public results to be ranked (#157); Qwen2.5 14B Instruct does not yet, so treat this comparison as directional.

- Canonical page: https://noometry.com/compare/deepseek-v3-vs-qwen2-5-14b-instruct
- Last updated: 2026-10-10
- Shared benchmarks: 3

## Summary

- They share 3 benchmarks with published results for both. DeepSeek-V3 scores higher in 1 category and Qwen2.5 14B Instruct in 0 categories; one gap is clear of the uncertainty.
- The widest gap is in coding, where DeepSeek-V3 leads 41.9 to 37.1.
- The biggest single-benchmark swing is BigCodeBench Instruct: 50% for DeepSeek-V3 and 39.8% for Qwen2.5 14B Instruct.
- DeepSeek-V3 is cheaper at $0.24 / $0.90 per million input/output tokens, against $0.35 / $1.40 for Qwen2.5 14B Instruct.
- DeepSeek-V3 accepts more context: 164K tokens versus 131K.

## Snapshot

| | DeepSeek-V3 | Qwen2.5 14B Instruct |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 39.2 | 38.2 |
| Rank | 157 | — |
| Context | 164K | 131K |
| Input $/M | $0.24 | $0.35 |
| Output $/M | $0.90 | $1.40 |
| Weights | Open | Open |

## Coding

- DeepSeek-V3: 41.9 (#99)
- Qwen2.5 14B Instruct: 37.1 (#187)

| Benchmark | DeepSeek-V3 | Qwen2.5 14B Instruct |
|---|---|---|
| BigCodeBench Instruct | 50% | 39.8% |
| BigCodeBench Complete | 62.2% | 52.2% |
| Aider Polyglot | 55.1% | — |
| SciCode | 35.8% | — |
| WeirdML | 36.1% | — |
| LiveBench Coding | 70.9% | — |
| LMArena Coding | 1368 | — |
| HumanEval+ | 86.6% | — |
| MBPP+ | 73% | — |

## Agentic & Tool Use

- DeepSeek-V3: —
- Qwen2.5 14B Instruct: —

| Benchmark | DeepSeek-V3 | Qwen2.5 14B Instruct |
|---|---|---|
| METR Time Horizons | 49.6% | — |

## Reasoning

- DeepSeek-V3: 20.0 (#226)
- Qwen2.5 14B Instruct: —

| Benchmark | DeepSeek-V3 | Qwen2.5 14B Instruct |
|---|---|---|
| SimpleBench | 27.2% | — |
| Kagi LLM Benchmark | 52.3% | — |
| CritPt | 0% | — |
| LiveBench Reasoning | 65.8% | — |
| LMArena Hard Prompts | 1365 | — |
| DTBench | 64.8% | — |
| LiveBench Data Analysis | 60.9% | — |
| LMCA | 15.5% | — |
| BIG-Bench Hard | 87.5% | — |
| Epoch Capabilities Index | 135.95 | — |
| ForecastBench | 59.1 | — |
| HellaSwag | 88.9% | — |
| LiveBench | 66.9% | — |
| PIQA | 84.7% | — |
| WinoGrande | 85.2% | — |

## Math

- DeepSeek-V3: 30.4 (#208)
- Qwen2.5 14B Instruct: —

| Benchmark | DeepSeek-V3 | Qwen2.5 14B Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 37.8% | — |
| Omni-MATH | 40.3% | — |
| LiveBench Math | 73.5% | — |
| LMArena Math | 1373 | — |
| MATH Level 5 | 75.5% | — |
| FrontierMath (Feb 2025 set) | 1.7% | — |

## Knowledge

- DeepSeek-V3: 37.5 (#147)
- Qwen2.5 14B Instruct: —

| Benchmark | DeepSeek-V3 | Qwen2.5 14B Instruct |
|---|---|---|
| MMLU | 87.2% | 79.9% |
| GPQA Diamond | 67.6% | — |
| MMLU-Pro | 72.3% | — |
| Confabulations | 26.1% | — |
| Vectara Hallucination Rate | 6.1% | — |
| GPQA (HELM) | 53.8% | — |
| LMArena Expert | 1352 | — |
| ARC (AI2) Challenge | 95.3% | — |
| TriviaQA | 82.9% | — |

## Multilingual

- DeepSeek-V3: 48.8 (#138)
- Qwen2.5 14B Instruct: —

| Benchmark | DeepSeek-V3 | Qwen2.5 14B Instruct |
|---|---|---|
| LMArena Non-English | 1358 | — |
| LMArena Chinese | 1390 | — |
| LMArena French | 1386 | — |
| LMArena German | 1374 | — |
| LMArena Japanese | 1333 | — |
| LMArena Korean | 1321 | — |
| LMArena Russian | 1372 | — |
| LMArena Spanish | 1358 | — |

## Instruction Following

- DeepSeek-V3: 73.0 (#125)
- Qwen2.5 14B Instruct: —

| Benchmark | DeepSeek-V3 | Qwen2.5 14B Instruct |
|---|---|---|
| LiveBench Instruction Following | 81.5% | — |
| IFEval | 83.2% | — |
| LMArena Instruction Following | 1345 | — |

## Long Context

- DeepSeek-V3: 34.0 (#248)
- Qwen2.5 14B Instruct: —

| Benchmark | DeepSeek-V3 | Qwen2.5 14B Instruct |
|---|---|---|
| Fiction.LiveBench | 50% | — |
| LMArena Longer Query | 1353 | — |

## Writing & Preference

- DeepSeek-V3: 57.6 (#124)
- Qwen2.5 14B Instruct: —

| Benchmark | DeepSeek-V3 | Qwen2.5 14B Instruct |
|---|---|---|
| LMArena Text | 1375 | — |
| LMArena Creative Writing | 1364 | — |
| Short-Story Creative Writing | 77% | — |
| EQ-Bench Creative Writing | 1472 | — |
| WildBench | 83% | — |
| LMArena Multi-Turn | 1389 | — |
| LiveBench Language | 49.1% | — |

## FAQ

### Is DeepSeek-V3 better than Qwen2.5 14B Instruct?

DeepSeek-V3 has enough public results to be ranked (#157); Qwen2.5 14B Instruct does not yet, so treat this comparison as directional.

### Which is cheaper, DeepSeek-V3 or Qwen2.5 14B Instruct?

DeepSeek-V3 is cheaper. It lists at $0.24 per million input tokens and $0.90 per million output tokens; Qwen2.5 14B Instruct lists at $0.35 and $1.40.

### Is DeepSeek-V3 or Qwen2.5 14B Instruct better for coding?

DeepSeek-V3 scores higher on coding benchmarks: 41.9 versus 37.1 in the Noometry coding category.

### Which has the bigger context window?

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

### How many benchmarks do DeepSeek-V3 and Qwen2.5 14B Instruct share?

3 benchmarks have published results for both models. DeepSeek-V3 has 60 scored results on Noometry and Qwen2.5 14B Instruct has 3.