Noometry

Model comparison

Grok 4 vs Qwen3.8 27B

Grok 4 is the stronger model overall, scoring 47.8 to 45.3 on the Noometry Index.

Last verified . 19 shared benchmarks.

Grok 4 xAI

47.8

Rank #49 Confirmed

Qwen3.8 27B Alibaba (Qwen)

45.3

Rank #64 Confirmed

Summary

  • They share 19 benchmarks with published results for both. Grok 4 scores higher in 5 categories and Qwen3.8 27B in 4 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in long context, where Grok 4 leads 63.1 to 44.3.
  • Qwen3.8 27B has downloadable open weights; the other is API-only.

Side by side

Grok 4Qwen3.8 27B
ProviderxAIAlibaba (Qwen)
Noometry Index47.845.3
Released2025-07-092026-08-14
WeightsProprietaryOpen
Context window—262K
Max output—33K
Input $ / M tokens—$0.99
Output $ / M tokens—$1.49
Results tracked4828

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

Category by category

Coding Too close to call

Grok 4: 49.9 (#44), Qwen3.8 27B: 50.3 (#42)

BenchmarkGrok 4Qwen3.8 27B
LMArena Coding14081487
Aider Polyglot79.6%—
LMArena WebDev—1592
SciCode—46.6%
WeirdML45.7%—

Agentic & Tool Use Not comparable

Grok 4: 32.1 (#62), Qwen3.8 27B: —

BenchmarkGrok 4Qwen3.8 27B
Terminal-Bench27.2%—
Berkeley Function Calling Leaderboard63%—
GDPval21.1%—
Cybench43%—
DeepResearch Bench47.3%—
BALROG43.6%—
LMArena Search1142—
METR Time Horizons66.6%—

Reasoning Grok 4 leads

Grok 4: 36.9 (#59), Qwen3.8 27B: 31.5 (#71)

BenchmarkGrok 4Qwen3.8 27B
LMArena Hard Prompts14091459
Epoch Capabilities Index146.46149.38
ARC-AGI-216%—
SimpleBench60.5%—
Kagi LLM Benchmark73.6%—
NYT Connections (extended)—54.5%
ARC-AGI-166.7%—
CritPt—5.4%
Chess Puzzles28%—
DTBench—88%
LMCA—41.4%
Surface Evolver Bench—45%
ForecastBench60.9—

Math Grok 4 leads

Grok 4: 46.5 (#58), Qwen3.8 27B: 35.7 (#139)

BenchmarkGrok 4Qwen3.8 27B
LMArena Math14221453
OTIS Mock AIME 2024-202584%—
ProofBench—16%
Omni-MATH60.3%—
FrontierMath (Feb 2025 set)19.7%—
FrontierMath Tier 4 (v1)2.1%—

Knowledge Grok 4 leads

Grok 4: 53.7 (#50), Qwen3.8 27B: 40.9 (#105)

BenchmarkGrok 4Qwen3.8 27B
LMArena Expert14171482
GPQA Diamond87%—
MMLU-Pro85.1%—
Confabulations12.4%—
GPQA (HELM)72.7%—

Multimodal Qwen3.8 27B leads

Grok 4: 33.1 (#89), Qwen3.8 27B: 40.5 (#36)

BenchmarkGrok 4Qwen3.8 27B
LMArena Vision12111270
GeoBench45%—

Multilingual Qwen3.8 27B leads

Grok 4: 52.0 (#98), Qwen3.8 27B: 54.0 (#54)

BenchmarkGrok 4Qwen3.8 27B
LMArena Non-English14031431
LMArena Chinese14281504
LMArena French14181468
LMArena German14291435
LMArena Japanese13961392
LMArena Korean13761389
LMArena Russian14091421
LMArena Spanish14191444

Instruction Following Grok 4 leads

Grok 4: 79.4 (#4), Qwen3.8 27B: 75.8 (#54)

BenchmarkGrok 4Qwen3.8 27B
LMArena Instruction Following13871436
IFEval94.9%—

Long Context Grok 4 leads

Grok 4: 63.1 (#4), Qwen3.8 27B: 44.3 (#66)

BenchmarkGrok 4Qwen3.8 27B
LMArena Longer Query14091448
Fiction.LiveBench94.4%—

Writing & Preference Qwen3.8 27B leads

Grok 4: 58.7 (#111), Qwen3.8 27B: 65.9 (#41)

BenchmarkGrok 4Qwen3.8 27B
LMArena Text14121441
LMArena Creative Writing13971382
LMArena Multi-Turn14161443
Short-Story Creative Writing76.9%—
EQ-Bench Creative Writing—1671
WildBench79.7%—

Frequently asked questions

Is Grok 4 better than Qwen3.8 27B?

Grok 4 is the stronger model overall, scoring 47.8 to 45.3 on the Noometry Index.

Is Grok 4 or Qwen3.8 27B better for coding?

They score almost the same on coding (49.9 vs 50.3); test both on your own repository before choosing.

How many benchmarks do Grok 4 and Qwen3.8 27B share?

19 benchmarks have published results for both models. Grok 4 has 48 scored results on Noometry and Qwen3.8 27B has 28.

Related comparisons

Go deeper