# Gemma 3 12B vs Gemma 7B

> Gemma 3 12B is the stronger model overall, scoring 31.8 to 29.5 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gemma-3-12b-vs-gemma-7b
- Last updated: 2026-10-10
- Shared benchmarks: 12

## Summary

- They share 12 benchmarks with published results for both. Gemma 3 12B scores higher in 5 categories and Gemma 7B in 3 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in multilingual, where Gemma 3 12B leads 45.9 to 25.3.

## Snapshot

| | Gemma 3 12B | Gemma 7B |
|---|---|---|
| Provider | Google | Google |
| Noometry Index | 31.8 | 29.5 |
| Rank | 255 | 295 |
| Context | 131K | — |
| Input $/M | $0.05 | — |
| Output $/M | $0.15 | — |
| Weights | Open | Open |

## Coding

- Gemma 3 12B: 31.2 (#276)
- Gemma 7B: 29.9 (#290)

| Benchmark | Gemma 3 12B | Gemma 7B |
|---|---|---|
| LMArena Coding | 1281 | 1048 |
| SciCode | 17.4% | — |
| HumanEval+ | — | 28.7% |
| MBPP+ | — | 43.4% |

## Agentic & Tool Use

- Gemma 3 12B: 25.3 (#101)
- Gemma 7B: —

| Benchmark | Gemma 3 12B | Gemma 7B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 30.4% | — |

## Reasoning

- Gemma 3 12B: 15.2 (#307)
- Gemma 7B: 19.0 (#247)

| Benchmark | Gemma 3 12B | Gemma 7B |
|---|---|---|
| LMArena Hard Prompts | 1309 | 1042 |
| Epoch Capabilities Index | 123.49 | 111.74 |
| CritPt | 0% | — |
| Chess Puzzles | 0% | — |
| DTBench | 48.8% | — |
| LMCA | 4.5% | — |
| Adversarial NLI | — | 48.7% |
| BIG-Bench Hard | — | 55.1% |
| HellaSwag | — | 82.2% |
| PIQA | — | 81.2% |
| WinoGrande | — | 79% |

## Math

- Gemma 3 12B: 21.0 (#273)
- Gemma 7B: 28.7 (#225)

| Benchmark | Gemma 3 12B | Gemma 7B |
|---|---|---|
| LMArena Math | 1307 | 1066 |
| OTIS Mock AIME 2024-2025 | 16.7% | — |
| GSM8K | — | 46.4% |

## Knowledge

- Gemma 3 12B: 26.5 (#250)
- Gemma 7B: 26.8 (#248)

| Benchmark | Gemma 3 12B | Gemma 7B |
|---|---|---|
| LMArena Expert | 1248 | 1002 |
| GPQA Diamond | 39.5% | — |
| Vectara Hallucination Rate | 4.4% | — |
| ARC (AI2) Challenge | — | 78.3% |
| BoolQ | — | 83.2% |
| MMLU | — | 66.1% |
| OpenBookQA | — | 78.6% |
| TriviaQA | — | 72.3% |

## Multimodal

- Gemma 3 12B: —
- Gemma 7B: —

| Benchmark | Gemma 3 12B | Gemma 7B |
|---|---|---|
| MindCube | 46.7% | — |

## Multilingual

- Gemma 3 12B: 45.9 (#160)
- Gemma 7B: 25.3 (#282)

| Benchmark | Gemma 3 12B | Gemma 7B |
|---|---|---|
| LMArena Non-English | 1318 | 999 |
| LMArena Russian | 1335 | 993 |
| LMArena Chinese | — | 1035 |
| LMArena French | — | 1024 |
| LMArena German | 1370 | — |

## Instruction Following

- Gemma 3 12B: 68.8 (#179)
- Gemma 7B: 51.7 (#290)

| Benchmark | Gemma 3 12B | Gemma 7B |
|---|---|---|
| LMArena Instruction Following | 1299 | 1017 |

## Long Context

- Gemma 3 12B: 40.0 (#157)
- Gemma 7B: 31.1 (#277)

| Benchmark | Gemma 3 12B | Gemma 7B |
|---|---|---|
| LMArena Longer Query | 1317 | 1022 |

## Writing & Preference

- Gemma 3 12B: 47.7 (#205)
- Gemma 7B: 27.2 (#297)

| Benchmark | Gemma 3 12B | Gemma 7B |
|---|---|---|
| LMArena Text | 1334 | 1056 |
| LMArena Creative Writing | 1331 | 1024 |
| LMArena Multi-Turn | 1334 | 963 |
| EQ-Bench Creative Writing | 1126 | — |

## FAQ

### Is Gemma 3 12B better than Gemma 7B?

Gemma 3 12B is the stronger model overall, scoring 31.8 to 29.5 on the Noometry Index.

### Is Gemma 3 12B or Gemma 7B better for coding?

Gemma 3 12B scores higher on coding benchmarks: 31.2 versus 29.9 in the Noometry coding category.

### How many benchmarks do Gemma 3 12B and Gemma 7B share?

12 benchmarks have published results for both models. Gemma 3 12B has 24 scored results on Noometry and Gemma 7B has 27.