Don't Overthink It: A Survey of Efficient R1-style Large Reasoning Models

2025-08-04

AI Summary

To address the “overthinking” problem in Large Reasoning Models (LRMs)—characterized by redundant reasoning, reduced efficiency, and degraded accuracy due to excessively long chains of thought—this work systematically surveys efficient inference methodologies for R1-style models and proposes, for the first time, a dual-path framework comprising single-model optimization and multi-model collaboration. Methodologically, it integrates reinforcement learning–driven self-reflection with long-chain reasoning, incorporating step compression and path pruning. Contributions include an open-source knowledge repository for continuous progress tracking and the distillation of reusable efficient reasoning paradigms. Experiments demonstrate that the framework significantly reduces average reasoning steps and latency while preserving answer accuracy, offering a systematic solution toward lightweight and practical deployment of LRMs.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Knowledge Representation and Reasoning: Computational Complexity of Reasoning
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Recently, Large Reasoning Models (LRMs) have gradually become a research hotspot due to their outstanding performance in handling complex tasks. Among them, DeepSeek R1 has garnered significant attention for its exceptional performance and open-source nature, driving advancements in the research of R1-style LRMs. Unlike traditional Large Language Models (LLMs), these models enhance logical deduction and decision-making capabilities during reasoning by incorporating mechanisms such as long chain-of-thought and self-reflection through reinforcement learning. However, with the widespread application of these models, the problem of overthinking has gradually emerged. Specifically, when generating answers, these models often construct excessively long reasoning chains with redundant or repetitive steps, which leads to reduced reasoning efficiency and may affect the accuracy of the final answer. To this end, various efficient reasoning methods have been proposed, aiming to reduce the length of reasoning paths without compromising model performance and reasoning capability. By reviewing the current research advancements in the field of efficient reasoning methods systematically, we categorize existing works into two main directions based on the lens of single-model optimization versus model collaboration: (1) Efficient Reasoning with Single Model, which focuses on improving the reasoning efficiency of individual models; and (2) Efficient Reasoning with Model Collaboration, which explores optimizing reasoning paths through collaboration among multiple models. Besides, we maintain a public GitHub repository that tracks the latest progress in efficient reasoning methods.

Problem

• Addresses overthinking in R1-style Large Reasoning Models
• Reduces redundant reasoning chains for efficiency
• Improves accuracy by optimizing reasoning paths

Innovation

• Long chain-of-thought for logical deduction
• Self-reflection via reinforcement learning
• Model collaboration for efficient reasoning