List your company

Mathematical Reasoning on MATH500 (Pass@1 and Pass@3)

85.6Pass@1UPFT (2025)
61.26467.58273.980.218Jan 15, 2026

Evaluation results

MethodLinks
UPFT (2025)Backbone=Qwen3-8B2026.0185.692.4
P-ALIGNBackbone=Qwen3-8B2026.0184.892.8
SFT (Long-CoT)Backbone=Qwen3-8B2026.0184.490.8
Zero-ShotBackbone=Qwen3-8B2026.0183.691.2
P-ALIGNBackbone=Qwen2.5-7B-Instruct2026.0175.885.2
SFT (Long-CoT)Backbone=Qwen2.5-7B-Instruct2026.0175.686.6
UPFT (2025)Backbone=Qwen2.5-7B-Instruct2026.0174.884.6
SFT (Label)Backbone=Qwen3-8B2026.0173.284.6
Zero-ShotBackbone=Qwen2.5-7B-Instruct2026.0169.884.8
SFT (Label)Backbone=Qwen2.5-7B-Instruct2026.0162.271