# Grok-3 mini vs Grok 4

> Grok 4 is the stronger model overall, scoring 48.1 to 41.2 on the Noometry Index.

- Canonical page: https://noometry.com/compare/grok-3-mini-vs-grok-4
- Last updated: 2026-10-10
- Shared benchmarks: 34

## Summary

- They share 34 benchmarks with published results for both. Grok-3 mini scores higher in 0 categories and Grok 4 in 8 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4 leads 36.7 to 13.6.
- The biggest single-benchmark swing is ARC-AGI-1: 16.5% for Grok-3 mini and 66.7% for Grok 4.

## Snapshot

| | Grok-3 mini | Grok 4 |
|---|---|---|
| Provider | xAI | xAI |
| Noometry Index | 41.2 | 48.1 |
| Rank | 141 | 56 |
| Context | — | — |
| Input $/M | — | — |
| Output $/M | — | — |
| Weights | Proprietary | Proprietary |

## Coding

- Grok-3 mini: 40.8 (#131)
- Grok 4: 50.3 (#46)

| Benchmark | Grok-3 mini | Grok 4 |
|---|---|---|
| Aider Polyglot | 49.3% | 79.6% |
| WeirdML | 42.6% | 45.7% |
| LMArena Coding | 1379 | 1408 |

## Agentic & Tool Use

- Grok-3 mini: —
- Grok 4: 32.3 (#68)

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

## Reasoning

- Grok-3 mini: 13.6 (#334)
- Grok 4: 36.7 (#65)

| Benchmark | Grok-3 mini | Grok 4 |
|---|---|---|
| ARC-AGI-2 | 0.4% | 16% |
| Kagi LLM Benchmark | 61.3% | 73.6% |
| ARC-AGI-1 | 16.5% | 66.7% |
| LMArena Hard Prompts | 1375 | 1409 |
| Epoch Capabilities Index | 140.35 | 146.44 |
| SimpleBench | — | 60.5% |
| Chess Puzzles | — | 28% |
| ForecastBench | — | 60.9 |

## Math

- Grok-3 mini: 42.1 (#85)
- Grok 4: 48.4 (#64)

| Benchmark | Grok-3 mini | Grok 4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 77.8% | 84% |
| Omni-MATH | 31.8% | 60.3% |
| LMArena Math | 1386 | 1422 |
| FrontierMath (Feb 2025 set) | 5.9% | 19.7% |
| MATH Level 5 | 90.9% | — |
| FrontierMath Tier 4 (v1) | — | 2.1% |

## Knowledge

- Grok-3 mini: 46.4 (#81)
- Grok 4: 53.8 (#55)

| Benchmark | Grok-3 mini | Grok 4 |
|---|---|---|
| GPQA Diamond | 76.3% | 87% |
| MMLU-Pro | 79.9% | 85.1% |
| Confabulations | 10.8% | 12.4% |
| GPQA (HELM) | 67.5% | 72.7% |
| LMArena Expert | 1395 | 1415 |

## Multimodal

- Grok-3 mini: —
- Grok 4: 33.7 (#94)

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

## Multilingual

- Grok-3 mini: 48.1 (#145)
- Grok 4: 51.8 (#103)

| Benchmark | Grok-3 mini | Grok 4 |
|---|---|---|
| LMArena Non-English | 1352 | 1403 |
| LMArena Chinese | 1387 | 1427 |
| LMArena French | 1357 | 1418 |
| LMArena German | 1349 | 1429 |
| LMArena Japanese | 1342 | 1394 |
| LMArena Korean | 1335 | 1377 |
| LMArena Russian | 1353 | 1410 |
| LMArena Spanish | 1381 | 1420 |

## Instruction Following

- Grok-3 mini: 78.5 (#9)
- Grok 4: 79.2 (#5)

| Benchmark | Grok-3 mini | Grok 4 |
|---|---|---|
| IFEval | 95.1% | 94.9% |
| LMArena Instruction Following | 1357 | 1387 |

## Long Context

- Grok-3 mini: 41.0 (#147)
- Grok 4: 63.1 (#4)

| Benchmark | Grok-3 mini | Grok 4 |
|---|---|---|
| Fiction.LiveBench | 66.7% | 94.4% |
| LMArena Longer Query | 1372 | 1409 |

## Writing & Preference

- Grok-3 mini: 52.5 (#169)
- Grok 4: 58.5 (#116)

| Benchmark | Grok-3 mini | Grok 4 |
|---|---|---|
| LMArena Text | 1370 | 1411 |
| LMArena Creative Writing | 1342 | 1397 |
| Short-Story Creative Writing | 73.5% | 76.9% |
| WildBench | 65.1% | 79.7% |
| LMArena Multi-Turn | 1355 | 1416 |

## FAQ

### Is Grok-3 mini better than Grok 4?

Grok 4 is the stronger model overall, scoring 48.1 to 41.2 on the Noometry Index.

### Is Grok-3 mini or Grok 4 better for coding?

Grok 4 scores higher on coding benchmarks: 50.3 versus 40.8 in the Noometry coding category.

### How many benchmarks do Grok-3 mini and Grok 4 share?

34 benchmarks have published results for both models. Grok-3 mini has 35 scored results on Noometry and Grok 4 has 48.