# DeepSeek-V3.2-Speciale vs Gemma 2 27B

> DeepSeek-V3.2-Speciale is the stronger model overall, scoring 39.7 to 29.4 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v3-2-speciale-vs-gemma-2-27b
- Last updated: 2026-10-12
- Shared benchmarks: 0

## Summary

- The widest gap is in reasoning, where DeepSeek-V3.2-Speciale leads 32.9 to 15.3.
- Gemma 2 27B is cheaper at $0.65 / $0.65 per million input/output tokens, against $0.58 / $1.68 for DeepSeek-V3.2-Speciale.
- DeepSeek-V3.2-Speciale accepts more context: 128K tokens versus 8K.

## Snapshot

| | DeepSeek-V3.2-Speciale | Gemma 2 27B |
|---|---|---|
| Provider | DeepSeek | Google |
| Noometry Index | 39.7 | 29.4 |
| Rank | 162 | 312 |
| Context | 128K | 8K |
| Input $/M | $0.58 | $0.65 |
| Output $/M | $1.68 | $0.65 |
| Weights | Open | Open |

## Coding

- DeepSeek-V3.2-Speciale: 40.4 (#140)
- Gemma 2 27B: 34.1 (#246)

| Benchmark | DeepSeek-V3.2-Speciale | Gemma 2 27B |
|---|---|---|
| WeirdML | 46.7% | — |
| BigCodeBench Instruct | — | 42.8% |
| LiveBench Coding | — | 36% |
| LMArena Coding | — | 1211 |
| BigCodeBench Complete | — | 52.5% |

## Reasoning

- DeepSeek-V3.2-Speciale: 32.9 (#73)
- Gemma 2 27B: 15.3 (#315)

| Benchmark | DeepSeek-V3.2-Speciale | Gemma 2 27B |
|---|---|---|
| SimpleBench | 52.6% | — |
| LiveBench Reasoning | — | 28.1% |
| LMArena Hard Prompts | — | 1198 |
| DTBench | — | 48% |
| LiveBench Data Analysis | — | 47.9% |
| LMCA | — | 7.1% |
| Epoch Capabilities Index | — | 122.08 |
| LiveBench | — | 38.2% |

## Math

- DeepSeek-V3.2-Speciale: —
- Gemma 2 27B: 10.7 (#311)

| Benchmark | DeepSeek-V3.2-Speciale | Gemma 2 27B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 1.4% |
| LiveBench Math | — | 26.5% |
| LMArena Math | — | 1212 |
| MATH Level 5 | — | 27.9% |

## Knowledge

- DeepSeek-V3.2-Speciale: —
- Gemma 2 27B: 19.0 (#280)

| Benchmark | DeepSeek-V3.2-Speciale | Gemma 2 27B |
|---|---|---|
| GPQA Diamond | — | 36.5% |
| Confabulations | — | 27.1% |
| LMArena Expert | — | 1172 |
| MMLU | — | 75.7% |

## Multilingual

- DeepSeek-V3.2-Speciale: —
- Gemma 2 27B: 38.6 (#226)

| Benchmark | DeepSeek-V3.2-Speciale | Gemma 2 27B |
|---|---|---|
| LMArena Non-English | — | 1217 |
| LMArena Chinese | — | 1221 |
| LMArena French | — | 1247 |
| LMArena German | — | 1209 |
| LMArena Japanese | — | 1175 |
| LMArena Korean | — | 1174 |
| LMArena Russian | — | 1234 |
| LMArena Spanish | — | 1228 |

## Instruction Following

- DeepSeek-V3.2-Speciale: —
- Gemma 2 27B: 60.5 (#249)

| Benchmark | DeepSeek-V3.2-Speciale | Gemma 2 27B |
|---|---|---|
| LiveBench Instruction Following | — | 58.1% |
| LMArena Instruction Following | — | 1206 |

## Long Context

- DeepSeek-V3.2-Speciale: —
- Gemma 2 27B: 37.3 (#218)

| Benchmark | DeepSeek-V3.2-Speciale | Gemma 2 27B |
|---|---|---|
| LMArena Longer Query | — | 1231 |

## Writing & Preference

- DeepSeek-V3.2-Speciale: 46.0 (#222)
- Gemma 2 27B: 44.2 (#225)

| Benchmark | DeepSeek-V3.2-Speciale | Gemma 2 27B |
|---|---|---|
| LMArena Text | — | 1231 |
| LMArena Creative Writing | — | 1241 |
| EQ-Bench Creative Writing | 1276 | — |
| LMArena Multi-Turn | — | 1224 |
| LiveBench Language | — | 32.6% |

## FAQ

### Is DeepSeek-V3.2-Speciale better than Gemma 2 27B?

DeepSeek-V3.2-Speciale is the stronger model overall, scoring 39.7 to 29.4 on the Noometry Index.

### Which is cheaper, DeepSeek-V3.2-Speciale or Gemma 2 27B?

Gemma 2 27B is cheaper. It lists at $0.65 per million input tokens and $0.65 per million output tokens; DeepSeek-V3.2-Speciale lists at $0.58 and $1.68.

### Is DeepSeek-V3.2-Speciale or Gemma 2 27B better for coding?

DeepSeek-V3.2-Speciale scores higher on coding benchmarks: 40.4 versus 34.1 in the Noometry coding category.

### Which has the bigger context window?

DeepSeek-V3.2-Speciale does, with 128K tokens against 8K.

### How many benchmarks do DeepSeek-V3.2-Speciale and Gemma 2 27B share?

0 benchmarks have published results for both models. DeepSeek-V3.2-Speciale has 3 scored results on Noometry and Gemma 2 27B has 34.