# DeepSeek-R1 vs Gemma 2 9B

> DeepSeek-R1 is the stronger model overall, scoring 42.0 to 25.7 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-r1-vs-gemma-2-9b
- Last updated: 2026-10-10
- Shared benchmarks: 29

## Summary

- They share 29 benchmarks with published results for both. DeepSeek-R1 scores higher in 8 categories and Gemma 2 9B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-R1 leads 44.3 to 9.7.
- The biggest single-benchmark swing is MATH Level 5: 96.6% for DeepSeek-R1 and 21% for Gemma 2 9B.
- Gemma 2 9B has downloadable open weights; the other is API-only.

## Snapshot

| | DeepSeek-R1 | Gemma 2 9B |
|---|---|---|
| Provider | DeepSeek | Google |
| Noometry Index | 42.0 | 25.7 |
| Rank | 110 | 334 |
| Context | 164K | — |
| Input $/M | $0.50 | — |
| Output $/M | $2.15 | — |
| Weights | Proprietary | Open |

## Coding

- DeepSeek-R1: 45.9 (#63)
- Gemma 2 9B: 29.0 (#299)

| Benchmark | DeepSeek-R1 | Gemma 2 9B |
|---|---|---|
| LiveBench Coding | 66.7% | 22.5% |
| LMArena Coding | 1427 | 1174 |
| Aider Polyglot | 71.4% | — |
| SciCode | 35.7% | — |
| WeirdML | 41.6% | — |
| BigCodeBench Instruct | — | 34.7% |
| BigCodeBench Complete | — | 40.6% |
| ALE-Bench | 804.12 | — |
| AlgoTune | 1.7 | — |

## Agentic & Tool Use

- DeepSeek-R1: 30.4 (#68)
- Gemma 2 9B: —

| Benchmark | DeepSeek-R1 | Gemma 2 9B |
|---|---|---|
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| METR Time Horizons | 53.8% | — |

## Reasoning

- DeepSeek-R1: 18.7 (#253)
- Gemma 2 9B: 15.3 (#306)

| Benchmark | DeepSeek-R1 | Gemma 2 9B |
|---|---|---|
| LiveBench Reasoning | 83.2% | 15.2% |
| LMArena Hard Prompts | 1415 | 1172 |
| LiveBench Data Analysis | 69.8% | 36.4% |
| Epoch Capabilities Index | 141.3 | 119.78 |
| LiveBench | 71.6% | 28.7% |
| ARC-AGI-2 | 1.3% | — |
| SimpleBench | 40.8% | — |
| Kagi LLM Benchmark | 69.4% | — |
| ARC-AGI-1 | 21.2% | — |
| CritPt | 1.1% | — |
| ForecastBench | 60 | — |
| PIQA | — | 83.7% |

## Math

- DeepSeek-R1: 42.0 (#70)
- Gemma 2 9B: 9.3 (#312)

| Benchmark | DeepSeek-R1 | Gemma 2 9B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 0.6% |
| LiveBench Math | 80.7% | 19.8% |
| LMArena Math | 1401 | 1183 |
| MATH Level 5 | 96.6% | 21% |
| Omni-MATH | 42.4% | — |
| GSM8K | — | 84.9% |

## Knowledge

- DeepSeek-R1: 44.3 (#81)
- Gemma 2 9B: 9.7 (#300)

| Benchmark | DeepSeek-R1 | Gemma 2 9B |
|---|---|---|
| GPQA Diamond | 76.3% | 27.5% |
| LMArena Expert | 1394 | 1147 |
| MMLU-Pro | 79.3% | — |
| Confabulations | 12.7% | — |
| Vectara Hallucination Rate | 11.3% | — |
| GPQA (HELM) | 66.6% | — |
| BoolQ | — | 85.7% |
| MMLU | — | 72.1% |

## Multilingual

- DeepSeek-R1: 52.6 (#80)
- Gemma 2 9B: 36.8 (#233)

| Benchmark | DeepSeek-R1 | Gemma 2 9B |
|---|---|---|
| LMArena Non-English | 1412 | 1188 |
| LMArena Chinese | 1443 | 1185 |
| LMArena French | 1415 | 1189 |
| LMArena German | 1405 | 1186 |
| LMArena Japanese | 1394 | 1144 |
| LMArena Korean | 1359 | 1137 |
| LMArena Russian | 1423 | 1200 |
| LMArena Spanish | 1410 | 1200 |

## Instruction Following

- DeepSeek-R1: 72.2 (#138)
- Gemma 2 9B: 57.8 (#264)

| Benchmark | DeepSeek-R1 | Gemma 2 9B |
|---|---|---|
| LiveBench Instruction Following | 80.5% | 52.6% |
| LMArena Instruction Following | 1382 | 1178 |
| IFEval | 78.4% | — |

## Long Context

- DeepSeek-R1: 45.4 (#36)
- Gemma 2 9B: 36.3 (#228)

| Benchmark | DeepSeek-R1 | Gemma 2 9B |
|---|---|---|
| LMArena Longer Query | 1391 | 1197 |
| Fiction.LiveBench | 75% | — |

## Writing & Preference

- DeepSeek-R1: 61.5 (#84)
- Gemma 2 9B: 32.3 (#276)

| Benchmark | DeepSeek-R1 | Gemma 2 9B |
|---|---|---|
| LMArena Text | 1428 | 1207 |
| LMArena Creative Writing | 1406 | 1206 |
| EQ-Bench Creative Writing | 1500 | 841 |
| LMArena Multi-Turn | 1404 | 1193 |
| LiveBench Language | 48.5% | 25.5% |
| Short-Story Creative Writing | 83% | — |
| WildBench | 82.8% | — |

## FAQ

### Is DeepSeek-R1 better than Gemma 2 9B?

DeepSeek-R1 is the stronger model overall, scoring 42.0 to 25.7 on the Noometry Index.

### Is DeepSeek-R1 or Gemma 2 9B better for coding?

DeepSeek-R1 scores higher on coding benchmarks: 45.9 versus 29.0 in the Noometry coding category.

### How many benchmarks do DeepSeek-R1 and Gemma 2 9B share?

29 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Gemma 2 9B has 35.