# DeepSeek-R1 vs Qwen2.5-Max

> DeepSeek-R1 is the stronger model overall, scoring 42.0 to 40.1 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-r1-vs-qwen2-5-max
- Last updated: 2026-10-10
- Shared benchmarks: 27

## Summary

- They share 27 benchmarks with published results for both. DeepSeek-R1 scores higher in 7 categories and Qwen2.5-Max in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-R1 leads 44.3 to 34.8.
- The biggest single-benchmark swing is LiveBench Reasoning: 83.2% for DeepSeek-R1 and 51.4% for Qwen2.5-Max.

## Snapshot

| | DeepSeek-R1 | Qwen2.5-Max |
|---|---|---|
| Provider | DeepSeek | Alibaba (Qwen) |
| Noometry Index | 42.0 | 40.1 |
| Rank | 110 | 142 |
| Context | 164K | — |
| Input $/M | $0.50 | — |
| Output $/M | $2.15 | — |
| Weights | Proprietary | Proprietary |

## Coding

- DeepSeek-R1: 45.9 (#63)
- Qwen2.5-Max: 41.2 (#118)

| Benchmark | DeepSeek-R1 | Qwen2.5-Max |
|---|---|---|
| LiveBench Coding | 66.7% | 64.4% |
| LMArena Coding | 1427 | 1359 |
| Aider Polyglot | 71.4% | — |
| SciCode | 35.7% | — |
| WeirdML | 41.6% | — |
| ALE-Bench | 804.12 | — |
| AlgoTune | 1.7 | — |

## Agentic & Tool Use

- DeepSeek-R1: 30.4 (#68)
- Qwen2.5-Max: —

| Benchmark | DeepSeek-R1 | Qwen2.5-Max |
|---|---|---|
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| METR Time Horizons | 53.8% | — |

## Reasoning

- DeepSeek-R1: 18.7 (#253)
- Qwen2.5-Max: 24.6 (#144)

| Benchmark | DeepSeek-R1 | Qwen2.5-Max |
|---|---|---|
| LiveBench Reasoning | 83.2% | 51.4% |
| LMArena Hard Prompts | 1415 | 1360 |
| LiveBench Data Analysis | 69.8% | 67.9% |
| Epoch Capabilities Index | 141.3 | 132.54 |
| LiveBench | 71.6% | 62.3% |
| ARC-AGI-2 | 1.3% | — |
| SimpleBench | 40.8% | — |
| Kagi LLM Benchmark | 69.4% | — |
| ARC-AGI-1 | 21.2% | — |
| CritPt | 1.1% | — |
| ForecastBench | 60 | — |

## Math

- DeepSeek-R1: 42.0 (#70)
- Qwen2.5-Max: 34.3 (#162)

| Benchmark | DeepSeek-R1 | Qwen2.5-Max |
|---|---|---|
| LiveBench Math | 80.7% | 58.4% |
| LMArena Math | 1401 | 1369 |
| OTIS Mock AIME 2024-2025 | 66.4% | — |
| Omni-MATH | 42.4% | — |
| MATH Level 5 | 96.6% | — |

## Knowledge

- DeepSeek-R1: 44.3 (#81)
- Qwen2.5-Max: 34.8 (#181)

| Benchmark | DeepSeek-R1 | Qwen2.5-Max |
|---|---|---|
| Confabulations | 12.7% | 21.8% |
| LMArena Expert | 1394 | 1337 |
| GPQA Diamond | 76.3% | — |
| MMLU-Pro | 79.3% | — |
| Vectara Hallucination Rate | 11.3% | — |
| GPQA (HELM) | 66.6% | — |

## Multilingual

- DeepSeek-R1: 52.6 (#80)
- Qwen2.5-Max: 48.3 (#140)

| Benchmark | DeepSeek-R1 | Qwen2.5-Max |
|---|---|---|
| LMArena Non-English | 1412 | 1352 |
| LMArena Chinese | 1443 | 1382 |
| LMArena French | 1415 | 1396 |
| LMArena German | 1405 | 1350 |
| LMArena Japanese | 1394 | 1299 |
| LMArena Korean | 1359 | 1304 |
| LMArena Russian | 1423 | 1353 |
| LMArena Spanish | 1410 | 1377 |

## Instruction Following

- DeepSeek-R1: 72.2 (#138)
- Qwen2.5-Max: 71.5 (#147)

| Benchmark | DeepSeek-R1 | Qwen2.5-Max |
|---|---|---|
| LiveBench Instruction Following | 80.5% | 75.3% |
| LMArena Instruction Following | 1382 | 1335 |
| IFEval | 78.4% | — |

## Long Context

- DeepSeek-R1: 45.4 (#36)
- Qwen2.5-Max: 41.3 (#137)

| Benchmark | DeepSeek-R1 | Qwen2.5-Max |
|---|---|---|
| LMArena Longer Query | 1391 | 1358 |
| Fiction.LiveBench | 75% | — |

## Writing & Preference

- DeepSeek-R1: 61.5 (#84)
- Qwen2.5-Max: 55.6 (#141)

| Benchmark | DeepSeek-R1 | Qwen2.5-Max |
|---|---|---|
| LMArena Text | 1428 | 1367 |
| LMArena Creative Writing | 1406 | 1339 |
| Short-Story Creative Writing | 83% | 72.9% |
| LMArena Multi-Turn | 1404 | 1364 |
| LiveBench Language | 48.5% | 56.3% |
| EQ-Bench Creative Writing | 1500 | — |
| WildBench | 82.8% | — |

## FAQ

### Is DeepSeek-R1 better than Qwen2.5-Max?

DeepSeek-R1 is the stronger model overall, scoring 42.0 to 40.1 on the Noometry Index.

### Is DeepSeek-R1 or Qwen2.5-Max better for coding?

DeepSeek-R1 scores higher on coding benchmarks: 45.9 versus 41.2 in the Noometry coding category.

### How many benchmarks do DeepSeek-R1 and Qwen2.5-Max share?

27 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Qwen2.5-Max has 27.