Tag
Researchers release SU-01, a 30B-A3B reasoning model achieving gold-medal-level performance on physics and math Olympiad problems using a unified scaling recipe for proof search.
This paper presents a simple and unified recipe combining supervised fine-tuning, two-stage reinforcement learning, and test-time scaling to train a reasoning model (SU-01) that achieves gold-medal-level performance on International Mathematical and Physics Olympiad problems.