Noometry

Model comparison

Grok 4.6 vs Qwen2.5-VL 72B Instruct

Grok 4.6 is the stronger model overall, scoring 56.9 to 29.9 on the Noometry Index.

Last verified . 1 shared benchmarks.

Grok 4.6 xAI

56.9

Rank #21 Confirmed

Qwen2.5-VL 72B Instruct Alibaba (Qwen)

29.9

Rank #302 Confirmed

Summary

  • They share 1 benchmark with published results for both. Grok 4.6 scores higher in 3 categories and Qwen2.5-VL 72B Instruct in 0 categories; 3 gaps are clear of the uncertainty.
  • The widest gap is in reasoning, where Grok 4.6 leads 61.4 to 20.7.
  • Grok 4.6 is cheaper at $2 / $6 per million input/output tokens, against $2.80 / $8.40 for Qwen2.5-VL 72B Instruct.
  • Grok 4.6 accepts more context: 500K tokens versus 131K.
  • Qwen2.5-VL 72B Instruct has downloadable open weights; the other is API-only.

Side by side

Grok 4.6Qwen2.5-VL 72B Instruct
ProviderxAIAlibaba (Qwen)
Noometry Index56.929.9
Released2026-08-122024-09
WeightsProprietaryOpen
Context window500K131K
Max output500K8K
Input $ / M tokens$2$2.80
Output $ / M tokens$6$8.40
Results tracked496

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding Not comparable

Grok 4.6: 58.5 (#16), Qwen2.5-VL 72B Instruct: —

BenchmarkGrok 4.6Qwen2.5-VL 72B Instruct
DeepSWE67.5%—
FrontierCode48%—
CursorBench41.4%—
LMArena WebDev1617—
FrontierSWE25.3%—
SciCode56.5%—
WeirdML67.3%—
LMArena Coding1465—
ALE-Bench1,508—

Agentic & Tool Use Grok 4.6 leads

Grok 4.6: 39.4 (#27), Qwen2.5-VL 72B Instruct: 18.6 (#144)

BenchmarkGrok 4.6Qwen2.5-VL 72B Instruct
APEX-Agents65.3%—
OSWorld—5%
GDP.pdf17.2%—
Vending-Bench 29,047—

Reasoning Grok 4.6 leads

Grok 4.6: 61.4 (#20), Qwen2.5-VL 72B Instruct: 20.7 (#233)

BenchmarkGrok 4.6Qwen2.5-VL 72B Instruct
ARC-AGI-267.1%—
SimpleBench75.9%—
Kagi LLM Benchmark—36%
NYT Connections (extended)80%—
ARC-AGI-187.5%—
CritPt19.7%—
Chess Puzzles40%—
EBR-Bench30.5%—
LMArena Hard Prompts1447—
Mystery Game Puzzles34%—
DTBench97.3%—
LMCA48.5%—
Epoch Capabilities Index156.44—

Math Not comparable

Grok 4.6: 67.0 (#24), Qwen2.5-VL 72B Instruct: —

BenchmarkGrok 4.6Qwen2.5-VL 72B Instruct
FrontierMath (Tiers 1-3)66%—
FrontierMath Tier 431.7%—
OTIS Mock AIME 2024-202599.2%—
ProofBench51%—
LMArena Math1423—

Knowledge Not comparable

Grok 4.6: 63.3 (#20), Qwen2.5-VL 72B Instruct: —

BenchmarkGrok 4.6Qwen2.5-VL 72B Instruct
GPQA Diamond94%—
SimpleQA Verified49.3%—
LMArena Expert1467—

Multimodal Grok 4.6 leads

Grok 4.6: 43.6 (#23), Qwen2.5-VL 72B Instruct: 33.5 (#97)

BenchmarkGrok 4.6Qwen2.5-VL 72B Instruct
LMArena Vision12631107
Video-MME—73.5%
GeoBench—62%
Blueprint-Bench 233.2%—
Furniture Assembly40%—
LMArena Document1452—
SpatialViz-Bench—33.3%

Multilingual Not comparable

Grok 4.6: 53.0 (#74), Qwen2.5-VL 72B Instruct: —

BenchmarkGrok 4.6Qwen2.5-VL 72B Instruct
LMArena Non-English1420—
LMArena Chinese1480—
LMArena French1461—
LMArena German1431—
LMArena Japanese1376—
LMArena Korean1397—
LMArena Russian1422—
LMArena Spanish1404—

Instruction Following Not comparable

Grok 4.6: 75.4 (#63), Qwen2.5-VL 72B Instruct: —

BenchmarkGrok 4.6Qwen2.5-VL 72B Instruct
LMArena Instruction Following1431—

Long Context Not comparable

Grok 4.6: 44.5 (#66), Qwen2.5-VL 72B Instruct: —

BenchmarkGrok 4.6Qwen2.5-VL 72B Instruct
LMArena Longer Query1454—

Writing & Preference Not comparable

Grok 4.6: 62.3 (#80), Qwen2.5-VL 72B Instruct: —

BenchmarkGrok 4.6Qwen2.5-VL 72B Instruct
LMArena Text1428—
LMArena Creative Writing1428—
LMArena Multi-Turn1425—

Frequently asked questions

Is Grok 4.6 better than Qwen2.5-VL 72B Instruct?

Grok 4.6 is the stronger model overall, scoring 56.9 to 29.9 on the Noometry Index.

Which is cheaper, Grok 4.6 or Qwen2.5-VL 72B Instruct?

Grok 4.6 is cheaper. It lists at $2 per million input tokens and $6 per million output tokens; Qwen2.5-VL 72B Instruct lists at $2.80 and $8.40.

Which has the bigger context window?

Grok 4.6 does, with 500K tokens against 131K.

How many benchmarks do Grok 4.6 and Qwen2.5-VL 72B Instruct share?

1 benchmark has published results for both models. Grok 4.6 has 49 scored results on Noometry and Qwen2.5-VL 72B Instruct has 6.

Related comparisons

Go deeper