Noometry

Model comparison

Grok 4 vs Kimi K2 Thinking Turbo

Grok 4 is the stronger model overall, scoring 48.1 to 45.8 on the Noometry Index.

Last verified . 20 shared benchmarks.

Grok 4 xAI

48.1

Rank #56 Confirmed

Kimi K2 Thinking Turbo Moonshot AI

45.8

Rank #70 Confirmed

Summary

  • They share 20 benchmarks with published results for both. Grok 4 scores higher in 7 categories and Kimi K2 Thinking Turbo in 1 category; 6 gaps are clear of the uncertainty.
  • The widest gap is in long context, where Grok 4 leads 63.1 to 43.2.
  • The biggest single-benchmark swing is Chess Puzzles: 28% for Grok 4 and 20% for Kimi K2 Thinking Turbo.
  • Kimi K2 Thinking Turbo has downloadable open weights; the other is API-only.

Side by side

Grok 4Kimi K2 Thinking Turbo
ProviderxAIMoonshot AI
Noometry Index48.145.8
Released2025-07-092025-11-06
WeightsProprietaryOpen
Context window——
Max output——
Input $ / M tokens——
Output $ / M tokens——
Results tracked4821

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

Category by category

Coding Grok 4 leads

Grok 4: 50.3 (#46), Kimi K2 Thinking Turbo: 38.4 (#178)

BenchmarkGrok 4Kimi K2 Thinking Turbo
LMArena Coding14081454
Aider Polyglot79.6%—
LMArena WebDev—1322
WeirdML45.7%—

Agentic & Tool Use Not comparable

Grok 4: 32.3 (#68), Kimi K2 Thinking Turbo: —

BenchmarkGrok 4Kimi K2 Thinking Turbo
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: 36.7 (#65), Kimi K2 Thinking Turbo: 32.5 (#75)

BenchmarkGrok 4Kimi K2 Thinking Turbo
Chess Puzzles28%20%
LMArena Hard Prompts14091428
ARC-AGI-216%—
SimpleBench60.5%—
Kagi LLM Benchmark73.6%—
ARC-AGI-166.7%—
Epoch Capabilities Index146.44—
ForecastBench60.9—

Math Too close to call

Grok 4: 48.4 (#64), Kimi K2 Thinking Turbo: 47.4 (#68)

BenchmarkGrok 4Kimi K2 Thinking Turbo
OTIS Mock AIME 2024-202584%83.1%
LMArena Math14221429
Omni-MATH60.3%—
FrontierMath (Feb 2025 set)19.7%—
FrontierMath Tier 4 (v1)2.1%—

Knowledge Grok 4 leads

Grok 4: 53.8 (#55), Kimi K2 Thinking Turbo: 50.9 (#69)

BenchmarkGrok 4Kimi K2 Thinking Turbo
GPQA Diamond87%84.2%
LMArena Expert14151439
MMLU-Pro85.1%—
Confabulations12.4%—
GPQA (HELM)72.7%—

Multimodal Not comparable

Grok 4: 33.7 (#94), Kimi K2 Thinking Turbo: —

BenchmarkGrok 4Kimi K2 Thinking Turbo
LMArena Vision1210—
GeoBench45%—

Multilingual Too close to call

Grok 4: 51.8 (#103), Kimi K2 Thinking Turbo: 51.4 (#109)

BenchmarkGrok 4Kimi K2 Thinking Turbo
LMArena Non-English14031398
LMArena Chinese14271456
LMArena French14181425
LMArena German14291390
LMArena Japanese13941357
LMArena Korean13771330
LMArena Russian14101391
LMArena Spanish14201406

Instruction Following Grok 4 leads

Grok 4: 79.2 (#5), Kimi K2 Thinking Turbo: 74.0 (#109)

BenchmarkGrok 4Kimi K2 Thinking Turbo
LMArena Instruction Following13871403
IFEval94.9%—

Long Context Grok 4 leads

Grok 4: 63.1 (#4), Kimi K2 Thinking Turbo: 43.2 (#102)

BenchmarkGrok 4Kimi K2 Thinking Turbo
LMArena Longer Query14091415
Fiction.LiveBench94.4%—

Writing & Preference Kimi K2 Thinking Turbo leads

Grok 4: 58.5 (#116), Kimi K2 Thinking Turbo: 60.0 (#104)

BenchmarkGrok 4Kimi K2 Thinking Turbo
LMArena Text14111415
LMArena Creative Writing13971374
LMArena Multi-Turn14161414
Short-Story Creative Writing76.9%—
WildBench79.7%—

Frequently asked questions

Is Grok 4 better than Kimi K2 Thinking Turbo?

Grok 4 is the stronger model overall, scoring 48.1 to 45.8 on the Noometry Index.

Is Grok 4 or Kimi K2 Thinking Turbo better for coding?

Grok 4 scores higher on coding benchmarks: 50.3 versus 38.4 in the Noometry coding category.

How many benchmarks do Grok 4 and Kimi K2 Thinking Turbo share?

20 benchmarks have published results for both models. Grok 4 has 48 scored results on Noometry and Kimi K2 Thinking Turbo has 21.

Related comparisons

Go deeper