# Grok-2 (Dec 2024) vs Qwen3.8 Max

> Qwen3.8 Max is the stronger model overall, scoring 56.8 to 33.7 on the Noometry Index.

- Canonical page: https://noometry.com/compare/grok-2-vs-qwen3-8-max
- Last updated: 2026-10-10
- Shared benchmarks: 21

## Summary

- They share 21 benchmarks with published results for both. Grok-2 (Dec 2024) scores higher in 0 categories and Qwen3.8 Max in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3.8 Max leads 73.2 to 20.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 11.5% for Grok-2 (Dec 2024) and 100% for Qwen3.8 Max.

## Snapshot

| | Grok-2 (Dec 2024) | Qwen3.8 Max |
|---|---|---|
| Provider | xAI | Alibaba (Qwen) |
| Noometry Index | 33.7 | 56.8 |
| Rank | 239 | 22 |
| Context | — | 1M |
| Input $/M | — | $2 |
| Output $/M | — | $6 |
| Weights | Proprietary | Proprietary |

## Coding

- Grok-2 (Dec 2024): 33.3 (#258)
- Qwen3.8 Max: 53.5 (#29)

| Benchmark | Grok-2 (Dec 2024) | Qwen3.8 Max |
|---|---|---|
| LMArena Coding | 1287 | 1502 |
| DeepSWE | — | 57.5% |
| LMArena WebDev | — | 1674 |
| FrontierSWE | — | 17.8% |
| SciCode | — | 53.2% |
| WeirdML | 22.2% | — |
| LiveBench Coding | 46.4% | — |

## Agentic & Tool Use

- Grok-2 (Dec 2024): —
- Qwen3.8 Max: 45.4 (#14)

| Benchmark | Grok-2 (Dec 2024) | Qwen3.8 Max |
|---|---|---|
| APEX-Agents | — | 63.3% |
| τ²-bench Banking | — | 55.1% |
| GDP.pdf | — | 23.2% |

## Reasoning

- Grok-2 (Dec 2024): 16.9 (#299)
- Qwen3.8 Max: 54.4 (#26)

| Benchmark | Grok-2 (Dec 2024) | Qwen3.8 Max |
|---|---|---|
| LMArena Hard Prompts | 1272 | 1496 |
| DTBench | 65.2% | 92% |
| Epoch Capabilities Index | 130.48 | 156.41 |
| SimpleBench | 22.7% | — |
| NYT Connections (extended) | — | 88.3% |
| CritPt | — | 20% |
| Chess Puzzles | — | 40% |
| LiveBench Reasoning | 54.8% | — |
| Mystery Game Puzzles | — | 38% |
| LiveBench Data Analysis | 54.5% | — |
| LMCA | — | 46.2% |
| LiveBench | 54.3% | — |

## Math

- Grok-2 (Dec 2024): 20.8 (#284)
- Qwen3.8 Max: 73.2 (#20)

| Benchmark | Grok-2 (Dec 2024) | Qwen3.8 Max |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 11.5% | 100% |
| LMArena Math | 1283 | 1499 |
| FrontierMath (Tiers 1-3) | — | 74.7% |
| FrontierMath Tier 4 | — | 46.3% |
| ProofBench | — | 58% |
| LiveBench Math | 54.9% | — |
| MATH Level 5 | 63.5% | — |
| FrontierMath (Feb 2025 set) | 0.7% | — |

## Knowledge

- Grok-2 (Dec 2024): 29.8 (#233)
- Qwen3.8 Max: 61.7 (#27)

| Benchmark | Grok-2 (Dec 2024) | Qwen3.8 Max |
|---|---|---|
| GPQA Diamond | 53.8% | 92.7% |
| LMArena Expert | 1254 | 1507 |
| SimpleQA Verified | — | 47.3% |
| Confabulations | 20.1% | — |

## Multimodal

- Grok-2 (Dec 2024): —
- Qwen3.8 Max: 37.2 (#75)

| Benchmark | Grok-2 (Dec 2024) | Qwen3.8 Max |
|---|---|---|
| LMArena Vision | — | 1314 |
| Furniture Assembly | — | 20% |

## Multilingual

- Grok-2 (Dec 2024): 43.1 (#188)
- Qwen3.8 Max: 56.7 (#18)

| Benchmark | Grok-2 (Dec 2024) | Qwen3.8 Max |
|---|---|---|
| LMArena Non-English | 1282 | 1472 |
| LMArena Chinese | 1289 | 1538 |
| LMArena French | 1318 | 1503 |
| LMArena German | 1287 | 1483 |
| LMArena Japanese | 1244 | 1467 |
| LMArena Korean | 1237 | 1461 |
| LMArena Russian | 1286 | 1481 |
| LMArena Spanish | 1281 | 1492 |

## Instruction Following

- Grok-2 (Dec 2024): 66.9 (#202)
- Qwen3.8 Max: 77.6 (#17)

| Benchmark | Grok-2 (Dec 2024) | Qwen3.8 Max |
|---|---|---|
| LMArena Instruction Following | 1270 | 1479 |
| LiveBench Instruction Following | 69.6% | — |

## Long Context

- Grok-2 (Dec 2024): 38.8 (#190)
- Qwen3.8 Max: 45.6 (#31)

| Benchmark | Grok-2 (Dec 2024) | Qwen3.8 Max |
|---|---|---|
| LMArena Longer Query | 1276 | 1489 |

## Writing & Preference

- Grok-2 (Dec 2024): 48.6 (#198)
- Qwen3.8 Max: 67.1 (#30)

| Benchmark | Grok-2 (Dec 2024) | Qwen3.8 Max |
|---|---|---|
| LMArena Text | 1305 | 1483 |
| LMArena Creative Writing | 1284 | 1479 |
| LMArena Multi-Turn | 1290 | 1489 |
| Short-Story Creative Writing | 63.6% | — |
| LiveBench Language | 45.6% | — |

## FAQ

### Is Grok-2 (Dec 2024) better than Qwen3.8 Max?

Qwen3.8 Max is the stronger model overall, scoring 56.8 to 33.7 on the Noometry Index.

### Is Grok-2 (Dec 2024) or Qwen3.8 Max better for coding?

Qwen3.8 Max scores higher on coding benchmarks: 53.5 versus 33.3 in the Noometry coding category.

### How many benchmarks do Grok-2 (Dec 2024) and Qwen3.8 Max share?

21 benchmarks have published results for both models. Grok-2 (Dec 2024) has 34 scored results on Noometry and Qwen3.8 Max has 39.