# Grok 4 vs Hy4 preview

> Grok 4 is the stronger model overall, scoring 47.8 to 44.6 on the Noometry Index.

- Canonical page: https://noometry.com/compare/grok-4-vs-hy4-preview
- Last updated: 2026-10-10
- Shared benchmarks: 0

## Summary

- The widest gap is in math, where Hy4 preview leads 54.5 to 46.5.
- Hy4 preview has downloadable open weights; the other is API-only.

## Snapshot

| | Grok 4 | Hy4 preview |
|---|---|---|
| Provider | xAI | Tencent |
| Noometry Index | 47.8 | 44.6 |
| Rank | 49 | 73 |
| Context | — | 1.05M |
| Input $/M | — | $0.83 |
| Output $/M | — | $2.50 |
| Weights | Proprietary | Open |

## Coding

- Grok 4: 49.9 (#44)
- Hy4 preview: 51.4 (#36)

| Benchmark | Grok 4 | Hy4 preview |
|---|---|---|
| Aider Polyglot | 79.6% | — |
| LMArena WebDev | — | 1633 |
| WeirdML | 45.7% | — |
| LMArena Coding | 1408 | — |

## Agentic & Tool Use

- Grok 4: 32.1 (#62)
- Hy4 preview: —

| Benchmark | Grok 4 | Hy4 preview |
|---|---|---|
| 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 4: 36.9 (#59)
- Hy4 preview: 30.6 (#75)

| Benchmark | Grok 4 | Hy4 preview |
|---|---|---|
| ARC-AGI-2 | 16% | — |
| SimpleBench | 60.5% | — |
| Kagi LLM Benchmark | 73.6% | — |
| NYT Connections (extended) | — | 68.2% |
| ARC-AGI-1 | 66.7% | — |
| Chess Puzzles | 28% | — |
| LMArena Hard Prompts | 1409 | — |
| Epoch Capabilities Index | 146.46 | — |
| ForecastBench | 60.9 | — |

## Math

- Grok 4: 46.5 (#58)
- Hy4 preview: 54.5 (#37)

| Benchmark | Grok 4 | Hy4 preview |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 84% | — |
| ProofBench | — | 75% |
| Omni-MATH | 60.3% | — |
| LMArena Math | 1422 | — |
| FrontierMath (Feb 2025 set) | 19.7% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |

## Knowledge

- Grok 4: 53.7 (#50)
- Hy4 preview: —

| Benchmark | Grok 4 | Hy4 preview |
|---|---|---|
| GPQA Diamond | 87% | — |
| MMLU-Pro | 85.1% | — |
| Confabulations | 12.4% | — |
| GPQA (HELM) | 72.7% | — |
| LMArena Expert | 1417 | — |

## Multimodal

- Grok 4: 33.1 (#89)
- Hy4 preview: —

| Benchmark | Grok 4 | Hy4 preview |
|---|---|---|
| LMArena Vision | 1211 | — |
| GeoBench | 45% | — |

## Multilingual

- Grok 4: 52.0 (#98)
- Hy4 preview: —

| Benchmark | Grok 4 | Hy4 preview |
|---|---|---|
| LMArena Non-English | 1403 | — |
| LMArena Chinese | 1428 | — |
| LMArena French | 1418 | — |
| LMArena German | 1429 | — |
| LMArena Japanese | 1396 | — |
| LMArena Korean | 1376 | — |
| LMArena Russian | 1409 | — |
| LMArena Spanish | 1419 | — |

## Instruction Following

- Grok 4: 79.4 (#4)
- Hy4 preview: —

| Benchmark | Grok 4 | Hy4 preview |
|---|---|---|
| IFEval | 94.9% | — |
| LMArena Instruction Following | 1387 | — |

## Long Context

- Grok 4: 63.1 (#4)
- Hy4 preview: —

| Benchmark | Grok 4 | Hy4 preview |
|---|---|---|
| Fiction.LiveBench | 94.4% | — |
| LMArena Longer Query | 1409 | — |

## Writing & Preference

- Grok 4: 58.7 (#111)
- Hy4 preview: —

| Benchmark | Grok 4 | Hy4 preview |
|---|---|---|
| LMArena Text | 1412 | — |
| LMArena Creative Writing | 1397 | — |
| Short-Story Creative Writing | 76.9% | — |
| WildBench | 79.7% | — |
| LMArena Multi-Turn | 1416 | — |

## FAQ

### Is Grok 4 better than Hy4 preview?

Grok 4 is the stronger model overall, scoring 47.8 to 44.6 on the Noometry Index.

### Is Grok 4 or Hy4 preview better for coding?

Hy4 preview scores higher on coding benchmarks: 51.4 versus 49.9 in the Noometry coding category.

### How many benchmarks do Grok 4 and Hy4 preview share?

0 benchmarks have published results for both models. Grok 4 has 48 scored results on Noometry and Hy4 preview has 3.