# Gemma 3 4B vs Grok 4

> Grok 4 is the stronger model overall, scoring 48.1 to 28.1 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gemma-3-4b-vs-grok-4
- Last updated: 2026-10-10
- Shared benchmarks: 18

## Summary

- They share 18 benchmarks with published results for both. Gemma 3 4B scores higher in 0 categories and Grok 4 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Grok 4 leads 53.8 to 11.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 7.5% for Gemma 3 4B and 84% for Grok 4.
- Gemma 3 4B has downloadable open weights; the other is API-only.

## Snapshot

| | Gemma 3 4B | Grok 4 |
|---|---|---|
| Provider | Google | xAI |
| Noometry Index | 28.1 | 48.1 |
| Rank | 326 | 56 |
| Context | 131K | — |
| Input $/M | $0.04 | — |
| Output $/M | $0.08 | — |
| Weights | Open | Proprietary |

## Coding

- Gemma 3 4B: 35.9 (#215)
- Grok 4: 50.3 (#46)

| Benchmark | Gemma 3 4B | Grok 4 |
|---|---|---|
| LMArena Coding | 1230 | 1408 |
| Aider Polyglot | — | 79.6% |
| WeirdML | — | 45.7% |

## Agentic & Tool Use

- Gemma 3 4B: 20.9 (#142)
- Grok 4: 32.3 (#68)

| Benchmark | Gemma 3 4B | Grok 4 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 19.6% | 63% |
| Terminal-Bench | — | 27.2% |
| GDPval | — | 21.1% |
| Cybench | — | 43% |
| DeepResearch Bench | — | 47.3% |
| BALROG | — | 43.6% |
| LMArena Search | — | 1142 |
| METR Time Horizons | — | 66.6% |

## Reasoning

- Gemma 3 4B: 13.2 (#335)
- Grok 4: 36.7 (#65)

| Benchmark | Gemma 3 4B | Grok 4 |
|---|---|---|
| Kagi LLM Benchmark | 25.2% | 73.6% |
| Chess Puzzles | 0% | 28% |
| LMArena Hard Prompts | 1253 | 1409 |
| Epoch Capabilities Index | 116.02 | 146.44 |
| ARC-AGI-2 | — | 16% |
| SimpleBench | — | 60.5% |
| ARC-AGI-1 | — | 66.7% |
| DTBench | 50.9% | — |
| LMCA | 2.8% | — |
| ForecastBench | — | 60.9 |

## Math

- Gemma 3 4B: 16.8 (#292)
- Grok 4: 48.4 (#64)

| Benchmark | Gemma 3 4B | Grok 4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 7.5% | 84% |
| LMArena Math | 1239 | 1422 |
| Omni-MATH | — | 60.3% |
| FrontierMath (Feb 2025 set) | — | 19.7% |
| FrontierMath Tier 4 (v1) | — | 2.1% |

## Knowledge

- Gemma 3 4B: 11.8 (#299)
- Grok 4: 53.8 (#55)

| Benchmark | Gemma 3 4B | Grok 4 |
|---|---|---|
| GPQA Diamond | 23.2% | 87% |
| LMArena Expert | 1223 | 1415 |
| MMLU-Pro | — | 85.1% |
| Confabulations | — | 12.4% |
| Vectara Hallucination Rate | 6.4% | — |
| GPQA (HELM) | — | 72.7% |

## Multimodal

- Gemma 3 4B: —
- Grok 4: 33.7 (#94)

| Benchmark | Gemma 3 4B | Grok 4 |
|---|---|---|
| LMArena Vision | — | 1210 |
| GeoBench | — | 45% |

## Multilingual

- Gemma 3 4B: 42.5 (#194)
- Grok 4: 51.8 (#103)

| Benchmark | Gemma 3 4B | Grok 4 |
|---|---|---|
| LMArena Non-English | 1273 | 1403 |
| LMArena German | 1281 | 1429 |
| LMArena Russian | 1294 | 1410 |
| LMArena Chinese | — | 1427 |
| LMArena French | — | 1418 |
| LMArena Japanese | — | 1394 |
| LMArena Korean | — | 1377 |
| LMArena Spanish | — | 1420 |

## Instruction Following

- Gemma 3 4B: 65.2 (#225)
- Grok 4: 79.2 (#5)

| Benchmark | Gemma 3 4B | Grok 4 |
|---|---|---|
| LMArena Instruction Following | 1239 | 1387 |
| IFEval | — | 94.9% |

## Long Context

- Gemma 3 4B: 38.7 (#194)
- Grok 4: 63.1 (#4)

| Benchmark | Gemma 3 4B | Grok 4 |
|---|---|---|
| LMArena Longer Query | 1273 | 1409 |
| Fiction.LiveBench | — | 94.4% |

## Writing & Preference

- Gemma 3 4B: 42.0 (#239)
- Grok 4: 58.5 (#116)

| Benchmark | Gemma 3 4B | Grok 4 |
|---|---|---|
| LMArena Text | 1291 | 1411 |
| LMArena Creative Writing | 1271 | 1397 |
| LMArena Multi-Turn | 1255 | 1416 |
| Short-Story Creative Writing | — | 76.9% |
| EQ-Bench Creative Writing | 1068 | — |
| WildBench | — | 79.7% |

## FAQ

### Is Gemma 3 4B better than Grok 4?

Grok 4 is the stronger model overall, scoring 48.1 to 28.1 on the Noometry Index.

### Is Gemma 3 4B or Grok 4 better for coding?

Grok 4 scores higher on coding benchmarks: 50.3 versus 35.9 in the Noometry coding category.

### How many benchmarks do Gemma 3 4B and Grok 4 share?

18 benchmarks have published results for both models. Gemma 3 4B has 22 scored results on Noometry and Grok 4 has 48.