List your company

Long-horizon Reasoning on Checkers Jumping

100Performance (n=11)Lookahead-Enhanced Atomic Decomposition
10.5633.785780.22Mar 6, 2026

Evaluation results

MethodLinks
Lookahead-Enhanced Atomic DecompositionModel=o4-mini, Voting=true, k=8, h=32026.031005280
Atomic DecompositionModel=o4-mini, Voting=true2026.0376160
Atomic DecompositionModel=o4-mini, Voting=false2026.031400