List your company

Scientific Reasoning on SciBench (Pass@10)

72.23Pass@10Entropy-Tree
56.931660.903364.87568.8467Jan 2, 2026

Evaluation results

MethodLinks
Entropy-TreeModel=Qwen2.5-32B-Instruct2026.0172.23
Entropy-TreeModel=Qwen2.5-14B-Instruct2026.0171.31
Multi-chainModel=Qwen2.5-14B-Instruct2026.0171.05
Multi-chainModel=Qwen2.5-32B-Instruct2026.0170.58
Entropy-TreeModel=Qwen2.5-7B-Instruct2026.0158.27
Multi-chainModel=Qwen2.5-7B-Instruct2026.0157.52