List your company

Reasoning on BBEH (ACC., LEN.)

75.8AccuracyCoT2-Meta
18.839233.627148.41563.2029Mar 19, 2026Mar 20, 2026Mar 22, 2026Mar 24, 2026Mar 26, 2026Mar 28, 2026Mar 30, 2026

Evaluation results

MethodLinks
CoT2-MetaStrategy=Ours (CoT2-Meta), Inference Budget=C=162026.0375.8—
ReST-MCTS*Strategy=ReST-MCTS*, Inference Budget=C=162026.0373.8—
Vanilla ToTStrategy=Vanilla ToT, Inference Budget=C=162026.0371.2—
Best-of-16Strategy=Best-of-16, Inference Budget=C=162026.0368.9—
Greedy CoTStrategy=Greedy CoT, Inference Budget=C=162026.0365.4—
DDPOModel=Qwen3-14b, theta=0.22026.0321.26,185
GRPOModel=Qwen3-14b2026.0321.036,851