Noometry

Model comparison

Grok-3 mini vs Kimi K2.5

Kimi K2.5 is the stronger model overall, scoring 48.1 to 41.2 on the Noometry Index.

Last verified . 26 shared benchmarks.

Grok-3 mini xAI

41.2

Rank #141 Confirmed

Kimi K2.5 Moonshot AI

48.1

Rank #57 Confirmed

Summary

  • They share 26 benchmarks with published results for both. Grok-3 mini scores higher in 1 category and Kimi K2.5 in 7 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in reasoning, where Kimi K2.5 leads 31.2 to 13.6.
  • The biggest single-benchmark swing is ARC-AGI-1: 16.5% for Grok-3 mini and 65.3% for Kimi K2.5.
  • Kimi K2.5 has downloadable open weights; the other is API-only.

Side by side

Grok-3 miniKimi K2.5
ProviderxAIMoonshot AI
Noometry Index41.248.1
Released2025-04-092026-01-27
WeightsProprietaryOpen
Context window—262K
Max output—262K
Input $ / M tokens—$0.45
Output $ / M tokens—$2.25
Results tracked3551

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

Category by category

Coding Kimi K2.5 leads

Grok-3 mini: 40.8 (#131), Kimi K2.5: 48.8 (#53)

BenchmarkGrok-3 miniKimi K2.5
WeirdML42.6%45.6%
LMArena Coding13791474
SWE-bench Verified—73.8%
SWE-bench Verified (bash only)—70.8%
Aider Polyglot49.3%—
LMArena WebDev—1437
SWE-bench Multilingual—67.3%
SciCode—49%
ALE-Bench—821.65

Agentic & Tool Use Not comparable

Grok-3 mini: —, Kimi K2.5: 34.2 (#48)

BenchmarkGrok-3 miniKimi K2.5
Terminal-Bench—43.2%
OSWorld—63.3%
Vending-Bench 2—1,198

Reasoning Kimi K2.5 leads

Grok-3 mini: 13.6 (#334), Kimi K2.5: 31.2 (#80)

BenchmarkGrok-3 miniKimi K2.5
ARC-AGI-20.4%11.8%
Kagi LLM Benchmark61.3%78.5%
ARC-AGI-116.5%65.3%
LMArena Hard Prompts13751453
Epoch Capabilities Index140.35148.03
SimpleBench—46.8%
NYT Connections (extended)—69.9%
CritPt—3.1%
Chess Puzzles—12%
EnigmaEval—3.4%
Thematic Generalization—69.4%

Math Kimi K2.5 leads

Grok-3 mini: 42.1 (#85), Kimi K2.5: 51.8 (#53)

BenchmarkGrok-3 miniKimi K2.5
OTIS Mock AIME 2024-202577.8%92.2%
LMArena Math13861470
FrontierMath (Feb 2025 set)5.9%27.9%
MathArena Final-Answer Competitions—62.3%
Omni-MATH31.8%—
MATH Level 590.9%—
FrontierMath Tier 4 (v1)—4.2%

Knowledge Kimi K2.5 leads

Grok-3 mini: 46.4 (#81), Kimi K2.5: 53.6 (#56)

BenchmarkGrok-3 miniKimi K2.5
GPQA Diamond76.3%87.6%
LMArena Expert13951466
Humanity's Last Exam—24.4%
SimpleQA Verified—34.3%
MMLU-Pro79.9%—
Confabulations10.8%—
Vectara Hallucination Rate—14.2%
GPQA (HELM)67.5%—

Multimodal Not comparable

Grok-3 mini: —, Kimi K2.5: 41.1 (#39)

BenchmarkGrok-3 miniKimi K2.5
LMArena Vision—1269
LMArena Document—1430

Multilingual Kimi K2.5 leads

Grok-3 mini: 48.1 (#145), Kimi K2.5: 53.9 (#53)

BenchmarkGrok-3 miniKimi K2.5
LMArena Non-English13521433
LMArena Chinese13871495
LMArena French13571454
LMArena German13491441
LMArena Japanese13421421
LMArena Korean13351410
LMArena Russian13531435
LMArena Spanish13811450

Instruction Following Grok-3 mini leads

Grok-3 mini: 78.5 (#9), Kimi K2.5: 75.3 (#64)

BenchmarkGrok-3 miniKimi K2.5
LMArena Instruction Following13571431
IFEval95.1%—

Long Context Kimi K2.5 leads

Grok-3 mini: 41.0 (#147), Kimi K2.5: 52.1 (#7)

BenchmarkGrok-3 miniKimi K2.5
Fiction.LiveBench66.7%86.1%
LMArena Longer Query13721445
CL-bench—19.3%
CL-bench Life—13.2%

Writing & Preference Kimi K2.5 leads

Grok-3 mini: 52.5 (#169), Kimi K2.5: 65.1 (#53)

BenchmarkGrok-3 miniKimi K2.5
LMArena Text13701445
LMArena Creative Writing13421423
LMArena Multi-Turn13551444
Short-Story Creative Writing73.5%—
EQ-Bench Creative Writing—1579
WildBench65.1%—

Frequently asked questions

Is Grok-3 mini better than Kimi K2.5?

Kimi K2.5 is the stronger model overall, scoring 48.1 to 41.2 on the Noometry Index.

Is Grok-3 mini or Kimi K2.5 better for coding?

Kimi K2.5 scores higher on coding benchmarks: 48.8 versus 40.8 in the Noometry coding category.

How many benchmarks do Grok-3 mini and Kimi K2.5 share?

26 benchmarks have published results for both models. Grok-3 mini has 35 scored results on Noometry and Kimi K2.5 has 51.

Related comparisons

Go deeper