CAT: Confidence-Adaptive Thinking for Efficient Reasoning of Large Reasoning Models
2026-07-01 • Computation and Language
Computation and LanguageArtificial Intelligence
AI summaryⓘ
The authors studied how large reasoning models sometimes spend too much time on easy questions, making them slower and less efficient. To fix this, they created a new method called Confidence-Adaptive Thinking (CAT) that uses the model's own confidence to decide how long it should think about a question. Their tests show that CAT helps the model give accurate answers faster by thinking less on simple problems and more on tough ones. This approach improves both speed and accuracy compared to older methods.
Large Reasoning ModelsChain-of-ThoughtModel ConfidenceInference EfficiencyDifficulty EstimationReasoning LengthToken OverheadAccuracy-Latency Tradeoff
Authors
Qizhi Jiang, Shuo Wang, Pei Ke, Yuhang Song, Ke Qin
Abstract
Large Reasoning Models (LRMs) have achieved remarkable success on complex tasks by leveraging long chain-of-thought (CoT) trajectories, yet they frequently exhibit overthinking on simple queries, resulting in significant token overhead and reduced inference efficiency. However, existing compression methods predominantly apply uniform length reduction or rely on coarse-grained difficulty estimation, often leading to performance degradation on difficult problems. To address this limitation, we propose Confidence-Adaptive Thinking (CAT), a framework that incorporates the model's intrinsic self-certainty signals as confidence into the preference optimization process, which autonomously modulates reasoning lengths based on problem difficulty. Experimental results show that CAT consistently outperforms state-of-the-art baselines on reasoning accuracy across multiple benchmarks on different base models. Our work enables LRMs to effectively compress confident responses while deliberating on uncertain ones, offering a potentially robust solution for balancing accuracy and latency in practical industrial scenarios.