1 paper
Ning Shang, Yifei Liu, Yi Zhu +12
We introduce rStar2-Agent, a 14B math reasoning model trained with agentic reinforcement learning to achieve frontier-level performance. Beyond current long CoT, the model demonstr…