Kimi K3 — Benchmarks

Benchmark scores for Kimi K3 aggregated from public leaderboards, with how it ranks among open models. See hardware requirements for what you need to run it.

Overall rank: #8 of 123 open modelscomposite 70.6/100 across 11 benchmarks in 4 categories · methodology

Coding

BenchmarkScoreOpen rankAll models
LiveBench Coding81.5#1 / 16#11 / 54
SciCode58.7%#1 / 50#5 / 155

Knowledge

BenchmarkScoreOpen rankAll models
SimpleQA50.6%#2 / 15#19 / 78

Math

BenchmarkScoreOpen rankAll models
LiveBench Math84.4#10 / 16#41 / 54
AIME 2024/202597.2%#2 / 74#23 / 270
ProofBench87.0%#1 / 21#5 / 64

Reasoning

BenchmarkScoreOpen rankAll models
LiveBench Reasoning90.7#1 / 16#5 / 54
GPQA Diamond93.1%#1 / 83#16 / 291
ARC-AGI94.5%#1 / 16#27 / 200
SimpleBench60.7%#2 / 23#29 / 101
CritPt23.4%#1 / 50#27 / 167

Scores aggregated from public benchmark sources (each linked from the benchmark pages). llmrun does not run these benchmarks.