# Grok 4.6 vs Qwen2-72B

> Grok 4.6 is the stronger model overall, scoring 56.8 to 29.5 on the Noometry Index.

- Canonical page: https://noometry.com/compare/grok-4-6-vs-qwen2-72b
- Last updated: 2026-10-10
- Shared benchmarks: 20

## Summary

- They share 20 benchmarks with published results for both. Grok 4.6 scores higher in 9 categories and Qwen2-72B in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Grok 4.6 leads 63.3 to 21.2.
- The biggest single-benchmark swing is WeirdML: 67.3% for Grok 4.6 and 11.3% for Qwen2-72B.
- Qwen2-72B has downloadable open weights; the other is API-only.

## Snapshot

| | Grok 4.6 | Qwen2-72B |
|---|---|---|
| Provider | xAI | Alibaba (Qwen) |
| Noometry Index | 56.8 | 29.5 |
| Rank | 19 | 294 |
| Context | 500K | — |
| Input $/M | $2 | — |
| Output $/M | $6 | — |
| Weights | Proprietary | Open |

## Coding

- Grok 4.6: 58.0 (#14)
- Qwen2-72B: 28.7 (#303)

| Benchmark | Grok 4.6 | Qwen2-72B |
|---|---|---|
| WeirdML | 67.3% | 11.3% |
| LMArena Coding | 1464 | 1196 |
| DeepSWE | 67.5% | — |
| FrontierCode | 48% | — |
| CursorBench | 41.4% | — |
| LMArena WebDev | 1620 | — |
| FrontierSWE | 25.3% | — |
| SciCode | 56.5% | — |
| BigCodeBench Instruct | — | 38.5% |
| BigCodeBench Complete | — | 54% |
| ALE-Bench | 1,508 | — |

## Agentic & Tool Use

- Grok 4.6: 39.4 (#25)
- Qwen2-72B: 16.9 (#138)

| Benchmark | Grok 4.6 | Qwen2-72B |
|---|---|---|
| APEX-Agents | 65.3% | — |
| TheAgentCompany | — | 1.1% |
| GDP.pdf | 17.2% | — |
| METR Time Horizons | — | 29.9% |
| Vending-Bench 2 | 9,047 | — |

## Reasoning

- Grok 4.6: 61.3 (#18)
- Qwen2-72B: 22.2 (#184)

| Benchmark | Grok 4.6 | Qwen2-72B |
|---|---|---|
| LMArena Hard Prompts | 1445 | 1191 |
| Epoch Capabilities Index | 156.48 | 125.27 |
| ARC-AGI-2 | 67.1% | — |
| SimpleBench | 75.9% | — |
| NYT Connections (extended) | 80% | — |
| ARC-AGI-1 | 87.5% | — |
| CritPt | 19.7% | — |
| Chess Puzzles | 40% | — |
| EBR-Bench | 30.5% | — |
| Mystery Game Puzzles | 34% | — |
| DTBench | 97.3% | — |
| LMCA | 48.5% | — |

## Math

- Grok 4.6: 66.8 (#19)
- Qwen2-72B: 27.9 (#232)

| Benchmark | Grok 4.6 | Qwen2-72B |
|---|---|---|
| LMArena Math | 1416 | 1235 |
| FrontierMath (Tiers 1-3) | 66% | — |
| FrontierMath Tier 4 | 31.7% | — |
| OTIS Mock AIME 2024-2025 | 99.2% | — |
| ProofBench | 51% | — |
| MATH Level 5 | — | 39.1% |

## Knowledge

- Grok 4.6: 63.3 (#18)
- Qwen2-72B: 21.2 (#270)

| Benchmark | Grok 4.6 | Qwen2-72B |
|---|---|---|
| GPQA Diamond | 94% | 40.8% |
| LMArena Expert | 1463 | 1171 |
| SimpleQA Verified | 49.3% | — |
| MMLU | — | 82.4% |

## Multimodal

- Grok 4.6: 43.9 (#18)
- Qwen2-72B: —

| Benchmark | Grok 4.6 | Qwen2-72B |
|---|---|---|
| LMArena Vision | 1265 | — |
| Blueprint-Bench 2 | 33.2% | — |
| Furniture Assembly | 40% | — |
| LMArena Document | 1452 | — |

## Multilingual

- Grok 4.6: 53.3 (#67)
- Qwen2-72B: 36.0 (#239)

| Benchmark | Grok 4.6 | Qwen2-72B |
|---|---|---|
| LMArena Non-English | 1422 | 1176 |
| LMArena Chinese | 1479 | 1240 |
| LMArena French | 1458 | 1169 |
| LMArena German | 1425 | 1151 |
| LMArena Japanese | 1383 | 1111 |
| LMArena Korean | 1395 | 1083 |
| LMArena Russian | 1427 | 1169 |
| LMArena Spanish | 1408 | 1168 |

## Instruction Following

- Grok 4.6: 75.6 (#59)
- Qwen2-72B: 62.0 (#236)

| Benchmark | Grok 4.6 | Qwen2-72B |
|---|---|---|
| LMArena Instruction Following | 1431 | 1181 |

## Long Context

- Grok 4.6: 44.4 (#62)
- Qwen2-72B: 36.1 (#230)

| Benchmark | Grok 4.6 | Qwen2-72B |
|---|---|---|
| LMArena Longer Query | 1453 | 1193 |

## Writing & Preference

- Grok 4.6: 62.5 (#77)
- Qwen2-72B: 41.0 (#236)

| Benchmark | Grok 4.6 | Qwen2-72B |
|---|---|---|
| LMArena Text | 1427 | 1203 |
| LMArena Creative Writing | 1430 | 1181 |
| LMArena Multi-Turn | 1422 | 1195 |

## FAQ

### Is Grok 4.6 better than Qwen2-72B?

Grok 4.6 is the stronger model overall, scoring 56.8 to 29.5 on the Noometry Index.

### Is Grok 4.6 or Qwen2-72B better for coding?

Grok 4.6 scores higher on coding benchmarks: 58.0 versus 28.7 in the Noometry coding category.

### How many benchmarks do Grok 4.6 and Qwen2-72B share?

20 benchmarks have published results for both models. Grok 4.6 has 49 scored results on Noometry and Qwen2-72B has 26.