Shengjia Zhang, Junjie Wu, Jiawei Chen, Changwang Zhang, Xingyu Lou, Wangchunshu Zhou, Sheng Zhou, Can Wang, Jun Wang
OThink-R1 is a method that reduces unnecessary reasoning in large reasoning models by switching between fast-thinking and slow-thinking modes, improving efficiency without losing accuracy.
In recent years, advanced reasoning models have become very good at solving complex problems by using a technique called chain-of-thought reasoning. However, these models sometimes use complex reasoning even for simple problems, which is inefficient. The researchers developed a method called OThink-R1 that can distinguish when complex reasoning is necessary and when a simpler approach will suffice. By switching between fast and slow thinking modes, OThink-R1 reduces unnecessary reasoning steps, making the models more efficient while maintaining their accuracy.
