# Gemma 2 9B vs Gemma 3 12B

> Gemma 3 12B is the stronger model overall, scoring 31.8 to 25.7 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gemma-2-9b-vs-gemma-3-12b
- Last updated: 2026-10-10
- Shared benchmarks: 16

## Summary

- They share 16 benchmarks with published results for both. Gemma 2 9B scores higher in 1 category and Gemma 3 12B in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Gemma 3 12B leads 26.5 to 9.7.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 0.6% for Gemma 2 9B and 16.7% for Gemma 3 12B.

## Snapshot

| | Gemma 2 9B | Gemma 3 12B |
|---|---|---|
| Provider | Google | Google |
| Noometry Index | 25.7 | 31.8 |
| Rank | 334 | 255 |
| Context | — | 131K |
| Input $/M | — | $0.05 |
| Output $/M | — | $0.15 |
| Weights | Open | Open |

## Coding

- Gemma 2 9B: 29.0 (#299)
- Gemma 3 12B: 31.2 (#276)

| Benchmark | Gemma 2 9B | Gemma 3 12B |
|---|---|---|
| LMArena Coding | 1174 | 1281 |
| SciCode | — | 17.4% |
| BigCodeBench Instruct | 34.7% | — |
| LiveBench Coding | 22.5% | — |
| BigCodeBench Complete | 40.6% | — |

## Agentic & Tool Use

- Gemma 2 9B: —
- Gemma 3 12B: 25.3 (#101)

| Benchmark | Gemma 2 9B | Gemma 3 12B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 30.4% |

## Reasoning

- Gemma 2 9B: 15.3 (#306)
- Gemma 3 12B: 15.2 (#307)

| Benchmark | Gemma 2 9B | Gemma 3 12B |
|---|---|---|
| LMArena Hard Prompts | 1172 | 1309 |
| Epoch Capabilities Index | 119.78 | 123.49 |
| CritPt | — | 0% |
| Chess Puzzles | — | 0% |
| LiveBench Reasoning | 15.2% | — |
| DTBench | — | 48.8% |
| LiveBench Data Analysis | 36.4% | — |
| LMCA | — | 4.5% |
| LiveBench | 28.7% | — |
| PIQA | 83.7% | — |

## Math

- Gemma 2 9B: 9.3 (#312)
- Gemma 3 12B: 21.0 (#273)

| Benchmark | Gemma 2 9B | Gemma 3 12B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 0.6% | 16.7% |
| LMArena Math | 1183 | 1307 |
| LiveBench Math | 19.8% | — |
| MATH Level 5 | 21% | — |
| GSM8K | 84.9% | — |

## Knowledge

- Gemma 2 9B: 9.7 (#300)
- Gemma 3 12B: 26.5 (#250)

| Benchmark | Gemma 2 9B | Gemma 3 12B |
|---|---|---|
| GPQA Diamond | 27.5% | 39.5% |
| LMArena Expert | 1147 | 1248 |
| Vectara Hallucination Rate | — | 4.4% |
| BoolQ | 85.7% | — |
| MMLU | 72.1% | — |

## Multimodal

- Gemma 2 9B: —
- Gemma 3 12B: —

| Benchmark | Gemma 2 9B | Gemma 3 12B |
|---|---|---|
| MindCube | — | 46.7% |

## Multilingual

- Gemma 2 9B: 36.8 (#233)
- Gemma 3 12B: 45.9 (#160)

| Benchmark | Gemma 2 9B | Gemma 3 12B |
|---|---|---|
| LMArena Non-English | 1188 | 1318 |
| LMArena German | 1186 | 1370 |
| LMArena Russian | 1200 | 1335 |
| LMArena Chinese | 1185 | — |
| LMArena French | 1189 | — |
| LMArena Japanese | 1144 | — |
| LMArena Korean | 1137 | — |
| LMArena Spanish | 1200 | — |

## Instruction Following

- Gemma 2 9B: 57.8 (#264)
- Gemma 3 12B: 68.8 (#179)

| Benchmark | Gemma 2 9B | Gemma 3 12B |
|---|---|---|
| LMArena Instruction Following | 1178 | 1299 |
| LiveBench Instruction Following | 52.6% | — |

## Long Context

- Gemma 2 9B: 36.3 (#228)
- Gemma 3 12B: 40.0 (#157)

| Benchmark | Gemma 2 9B | Gemma 3 12B |
|---|---|---|
| LMArena Longer Query | 1197 | 1317 |

## Writing & Preference

- Gemma 2 9B: 32.3 (#276)
- Gemma 3 12B: 47.7 (#205)

| Benchmark | Gemma 2 9B | Gemma 3 12B |
|---|---|---|
| LMArena Text | 1207 | 1334 |
| LMArena Creative Writing | 1206 | 1331 |
| EQ-Bench Creative Writing | 841 | 1126 |
| LMArena Multi-Turn | 1193 | 1334 |
| LiveBench Language | 25.5% | — |

## FAQ

### Is Gemma 2 9B better than Gemma 3 12B?

Gemma 3 12B is the stronger model overall, scoring 31.8 to 25.7 on the Noometry Index.

### Is Gemma 2 9B or Gemma 3 12B better for coding?

Gemma 3 12B scores higher on coding benchmarks: 31.2 versus 29.0 in the Noometry coding category.

### How many benchmarks do Gemma 2 9B and Gemma 3 12B share?

16 benchmarks have published results for both models. Gemma 2 9B has 35 scored results on Noometry and Gemma 3 12B has 24.