# DeepSeek LLM 67B vs Qwen2.5 72B Instruct

> Qwen2.5 72B Instruct is the stronger model overall, scoring 31.9 to 24.9 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-llm-67b-vs-qwen2-5-72b-instruct
- Last updated: 2026-10-12
- Shared benchmarks: 14

## Summary

- They share 14 benchmarks with published results for both. DeepSeek LLM 67B scores higher in 0 categories and Qwen2.5 72B Instruct in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Qwen2.5 72B Instruct leads 27.0 to 7.0.
- The biggest single-benchmark swing is MATH Level 5: 6.4% for DeepSeek LLM 67B and 63.2% for Qwen2.5 72B Instruct.

## Snapshot

| | DeepSeek LLM 67B | Qwen2.5 72B Instruct |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 24.9 | 31.9 |
| Rank | 347 | 267 |
| Context | — | 131K |
| Input $/M | — | $1.40 |
| Output $/M | — | $5.60 |
| Weights | Open | Open |

## Coding

- DeepSeek LLM 67B: 31.9 (#278)
- Qwen2.5 72B Instruct: 33.2 (#260)

| Benchmark | DeepSeek LLM 67B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Coding | 1096 | 1292 |
| WeirdML | — | 16% |
| BigCodeBench Instruct | — | 45.8% |
| BigCodeBench Complete | — | 55.9% |

## Agentic & Tool Use

- DeepSeek LLM 67B: —
- Qwen2.5 72B Instruct: 22.1 (#133)

| Benchmark | DeepSeek LLM 67B | Qwen2.5 72B Instruct |
|---|---|---|
| TheAgentCompany | — | 5.7% |
| BALROG | — | 16.2% |
| METR Time Horizons | — | 35.8% |

## Reasoning

- DeepSeek LLM 67B: 16.5 (#304)
- Qwen2.5 72B Instruct: 22.3 (#199)

| Benchmark | DeepSeek LLM 67B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Hard Prompts | 1070 | 1271 |
| Epoch Capabilities Index | 110.5 | 129 |
| Chess Puzzles | 0% | — |
| DTBench | — | 62.9% |
| LMCA | — | 13.4% |
| BIG-Bench Hard | — | 79.8% |
| ForecastBench | — | 57.5 |
| HellaSwag | — | 84.8% |
| PIQA | — | 82.6% |
| WinoGrande | — | 82.3% |

## Math

- DeepSeek LLM 67B: 8.7 (#324)
- Qwen2.5 72B Instruct: 19.3 (#287)

| Benchmark | DeepSeek LLM 67B | Qwen2.5 72B Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 0.8% | 8.1% |
| LMArena Math | 1108 | 1283 |
| MATH Level 5 | 6.4% | 63.2% |
| Omni-MATH | — | 33% |

## Knowledge

- DeepSeek LLM 67B: 7.0 (#313)
- Qwen2.5 72B Instruct: 27.0 (#253)

| Benchmark | DeepSeek LLM 67B | Qwen2.5 72B Instruct |
|---|---|---|
| GPQA Diamond | 24.6% | 49.1% |
| MMLU-Pro | — | 63.1% |
| Confabulations | — | 19.1% |
| GPQA (HELM) | — | 42.6% |
| LMArena Expert | — | 1245 |
| ARC (AI2) Challenge | — | 94.5% |
| MMLU | — | 85.3% |
| TriviaQA | — | 71.9% |

## Multilingual

- DeepSeek LLM 67B: 29.4 (#267)
- Qwen2.5 72B Instruct: 41.0 (#213)

| Benchmark | DeepSeek LLM 67B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Non-English | 1073 | 1252 |
| LMArena Chinese | 1132 | 1272 |
| LMArena French | — | 1280 |
| LMArena German | — | 1234 |
| LMArena Japanese | — | 1180 |
| LMArena Korean | — | 1188 |
| LMArena Russian | — | 1264 |
| LMArena Spanish | — | 1256 |

## Instruction Following

- DeepSeek LLM 67B: 55.4 (#277)
- Qwen2.5 72B Instruct: 65.5 (#221)

| Benchmark | DeepSeek LLM 67B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Instruction Following | 1079 | 1254 |
| IFEval | — | 80.6% |

## Long Context

- DeepSeek LLM 67B: 33.1 (#265)
- Qwen2.5 72B Instruct: 38.9 (#188)

| Benchmark | DeepSeek LLM 67B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Longer Query | 1092 | 1282 |

## Writing & Preference

- DeepSeek LLM 67B: 31.6 (#282)
- Qwen2.5 72B Instruct: 46.7 (#215)

| Benchmark | DeepSeek LLM 67B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Text | 1105 | 1269 |
| LMArena Creative Writing | 1067 | 1221 |
| LMArena Multi-Turn | 1082 | 1272 |
| WildBench | — | 80.2% |

## FAQ

### Is DeepSeek LLM 67B better than Qwen2.5 72B Instruct?

Qwen2.5 72B Instruct is the stronger model overall, scoring 31.9 to 24.9 on the Noometry Index.

### Is DeepSeek LLM 67B or Qwen2.5 72B Instruct better for coding?

Qwen2.5 72B Instruct scores higher on coding benchmarks: 33.2 versus 31.9 in the Noometry coding category.

### How many benchmarks do DeepSeek LLM 67B and Qwen2.5 72B Instruct share?

14 benchmarks have published results for both models. DeepSeek LLM 67B has 15 scored results on Noometry and Qwen2.5 72B Instruct has 43.