# Grok 3 vs Grok 4.6

> Grok 4.6 is the stronger model overall, scoring 56.8 to 39.6 on the Noometry Index.

- Canonical page: https://noometry.com/compare/grok-3-vs-grok-4-6
- Last updated: 2026-10-10
- Shared benchmarks: 24

## Summary

- They share 24 benchmarks with published results for both. Grok 3 scores higher in 0 categories and Grok 4.6 in 9 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4.6 leads 61.3 to 13.8.
- The biggest single-benchmark swing is ARC-AGI-1: 5.5% for Grok 3 and 87.5% for Grok 4.6.

## Snapshot

| | Grok 3 | Grok 4.6 |
|---|---|---|
| Provider | xAI | xAI |
| Noometry Index | 39.6 | 56.8 |
| Rank | 150 | 19 |
| Context | — | 500K |
| Input $/M | — | $2 |
| Output $/M | — | $6 |
| Weights | Proprietary | Proprietary |

## Coding

- Grok 3: 41.5 (#113)
- Grok 4.6: 58.0 (#14)

| Benchmark | Grok 3 | Grok 4.6 |
|---|---|---|
| WeirdML | 37.2% | 67.3% |
| LMArena Coding | 1432 | 1464 |
| DeepSWE | — | 67.5% |
| FrontierCode | — | 48% |
| Aider Polyglot | 53.3% | — |
| CursorBench | — | 41.4% |
| LMArena WebDev | — | 1620 |
| FrontierSWE | — | 25.3% |
| SciCode | — | 56.5% |
| ALE-Bench | — | 1,508 |

## Agentic & Tool Use

- Grok 3: 30.2 (#69)
- Grok 4.6: 39.4 (#25)

| Benchmark | Grok 3 | Grok 4.6 |
|---|---|---|
| APEX-Agents | — | 65.3% |
| BALROG | 29.5% | — |
| GDP.pdf | — | 17.2% |
| Vending-Bench 2 | — | 9,047 |

## Reasoning

- Grok 3: 13.8 (#322)
- Grok 4.6: 61.3 (#18)

| Benchmark | Grok 3 | Grok 4.6 |
|---|---|---|
| ARC-AGI-2 | 0% | 67.1% |
| SimpleBench | 36.1% | 75.9% |
| ARC-AGI-1 | 5.5% | 87.5% |
| LMArena Hard Prompts | 1433 | 1445 |
| Epoch Capabilities Index | 138.34 | 156.48 |
| Kagi LLM Benchmark | 61.3% | — |
| NYT Connections (extended) | — | 80% |
| CritPt | — | 19.7% |
| Chess Puzzles | — | 40% |
| EBR-Bench | — | 30.5% |
| Mystery Game Puzzles | — | 34% |
| DTBench | — | 97.3% |
| LMCA | — | 48.5% |

## Math

- Grok 3: 36.2 (#122)
- Grok 4.6: 66.8 (#19)

| Benchmark | Grok 3 | Grok 4.6 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 55.6% | 99.2% |
| LMArena Math | 1390 | 1416 |
| FrontierMath (Tiers 1-3) | — | 66% |
| FrontierMath Tier 4 | — | 31.7% |
| ProofBench | — | 51% |
| Omni-MATH | 46.4% | — |
| MATH Level 5 | 88.7% | — |
| FrontierMath (Feb 2025 set) | 3.8% | — |
| FrontierMath Tier 4 (v1) | 0% | — |

## Knowledge

- Grok 3: 46.0 (#76)
- Grok 4.6: 63.3 (#18)

| Benchmark | Grok 3 | Grok 4.6 |
|---|---|---|
| GPQA Diamond | 75.8% | 94% |
| LMArena Expert | 1420 | 1463 |
| SimpleQA Verified | — | 49.3% |
| MMLU-Pro | 78.8% | — |
| Confabulations | 14.2% | — |
| Vectara Hallucination Rate | 5.8% | — |
| GPQA (HELM) | 65% | — |

## Multimodal

- Grok 3: —
- Grok 4.6: 43.9 (#18)

| Benchmark | Grok 3 | Grok 4.6 |
|---|---|---|
| LMArena Vision | — | 1265 |
| Blueprint-Bench 2 | — | 33.2% |
| Furniture Assembly | — | 40% |
| LMArena Document | — | 1452 |

## Multilingual

- Grok 3: 52.5 (#81)
- Grok 4.6: 53.3 (#67)

| Benchmark | Grok 3 | Grok 4.6 |
|---|---|---|
| LMArena Non-English | 1411 | 1422 |
| LMArena Chinese | 1448 | 1479 |
| LMArena French | 1460 | 1458 |
| LMArena German | 1431 | 1425 |
| LMArena Japanese | 1389 | 1383 |
| LMArena Korean | 1374 | 1395 |
| LMArena Russian | 1416 | 1427 |
| LMArena Spanish | 1417 | 1408 |

## Instruction Following

- Grok 3: 75.2 (#69)
- Grok 4.6: 75.6 (#59)

| Benchmark | Grok 3 | Grok 4.6 |
|---|---|---|
| LMArena Instruction Following | 1409 | 1431 |
| IFEval | 88.4% | — |

## Long Context

- Grok 3: 38.7 (#187)
- Grok 4.6: 44.4 (#62)

| Benchmark | Grok 3 | Grok 4.6 |
|---|---|---|
| LMArena Longer Query | 1439 | 1453 |
| Fiction.LiveBench | 58.3% | — |

## Writing & Preference

- Grok 3: 56.0 (#136)
- Grok 4.6: 62.5 (#77)

| Benchmark | Grok 3 | Grok 4.6 |
|---|---|---|
| LMArena Text | 1426 | 1427 |
| LMArena Creative Writing | 1414 | 1430 |
| LMArena Multi-Turn | 1424 | 1422 |
| Short-Story Creative Writing | 76.4% | — |
| EQ-Bench Creative Writing | 1186 | — |
| WildBench | 84.9% | — |

## FAQ

### Is Grok 3 better than Grok 4.6?

Grok 4.6 is the stronger model overall, scoring 56.8 to 39.6 on the Noometry Index.

### Is Grok 3 or Grok 4.6 better for coding?

Grok 4.6 scores higher on coding benchmarks: 58.0 versus 41.5 in the Noometry coding category.

### How many benchmarks do Grok 3 and Grok 4.6 share?

24 benchmarks have published results for both models. Grok 3 has 40 scored results on Noometry and Grok 4.6 has 49.