Correct, Concise and Complete: Multi-stage Training For Adaptive Reasoning

The reasoning capabilities of large language models (LLMs) have improved substantially through increased test-time computation, typically in the form of intermediate tokens known as chain-of-thought (CoT). However, CoT often becomes unnecessarily long, increasing computation costs without improving accuracy and sometimes even degrading performance, a phenomenon known as "overthinking". We propose a multi-stage efficient reasoning method that combines supervised finetuning-via rejection sampling or reasoning trace reformatting-with reinforcement learning using an adaptive length penalty. We introduce a lightweight reward function that penalizes tokens generated after the first correct answer, encouraging the model to perform selfverification only when beneficial. We conduct a holistic evaluation across seven diverse reasoning tasks, analyzing the accuracy-response length trade-off. Our approach reduces response length by an average of 28% for 8B models and 40% for 32B models, while incurring only minor performance drops of 1.6 and 2.5 points, respectively. Despite its conceptual simplicity, it achieves a better trade-off than more complex state-of-the-art efficient reasoning methods, scoring 76.6 on the area under the Overthinking-Adjusted Accuracy curve (AUC OAA )-5 points above the base model and 2.5 points above the second-best approach.