综述如何让推理模型更快更省资源，三类方法全解析
Efficient Reasoning Models: A Survey
- 通过压缩、小型化和加速策略提升推理效率
- 实现短链条推理，模型体积减小，推理速度显著提升
- 适合关注大模型部署优化的研究者与工程师
推理模型通过生成长序列的思维链（Chain-of-Thoughts, CoTs）在复杂逻辑任务中表现卓越，但其逐词生成的‘慢思考’模式带来巨大计算开销。本文系统综述近期高效推理模型的研究进展，将其分为三大方向：(1) 短化——将冗长思维链压缩为简洁有效路径；(2) 小型化——通过知识蒸馏、模型压缩和强化学习等技术构建轻量级但具备强推理能力的语言模型；(3) 加速——设计高效的解码策略以加快推理过程。文中精选相关论文已整理至 GitHub 仓库：https://github.com/fscdc/Awesome-Efficient-Reasoning-Models。
原文摘要 · Abstract (English)
Reasoning models have demonstrated remarkable progress in solving complex and logic-intensive tasks by generating extended Chain-of-Thoughts (CoTs) prior to arriving at a final answer. Yet, the emergence of this "slow-thinking" paradigm, with numerous tokens generated in sequence, inevitably introduces substantial computational overhead. To this end, it highlights an urgent need for effective acceleration. This survey aims to provide a comprehensive overview of recent advances in efficient reasoning. It categorizes existing works into three key directions: (1) shorter - compressing lengthy CoTs into concise yet effective reasoning chains; (2) smaller - developing compact language models with strong reasoning capabilities through techniques such as knowledge distillation, other model compression techniques, and reinforcement learning; and (3) faster - designing efficient decoding strategies to accelerate inference of reasoning models. A curated collection of papers discussed in this survey is available in our GitHub repository: https://github.com/fscdc/Awesome-Efficient-Reasoning-Models.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。