Real-world applications of Large Reasoning Models (LRMs) often require reasoning about changing prompts or environments. In this work, we challenge the frozen world assumption and evaluate LRM robustness under two realistic dynamic scenarios: interruptions, which test the accuracy of model responses under budget-constrained outputs, and dynamic context, which tests model adaptation to in-flight changes. Across mathematics and programming benchmarks that require long-form reasoning, static evaluations consistently overestimate robustness: even state-of-the-art LRMs, which achieve high accuracy in static settings, can fail unpredictably when interrupted or exposed to changing context, with performance dropping by up to 60% when updates are introduced late in the reasoning process. Our analysis further reveals several novel failure modes, including reasoning leakage, where models fold the reasoning into their final answer when interrupted; panic, where under time pressure models abandon reasoning entirely and return incorrect answers; and self-doubt, where performance degrades when trying to incorporate updated information. Project Page: http://dynamic-lm.github.io/

翻译：大型推理模型（LRMs）在现实应用场景中常需基于动态变化的提示或环境进行推理。本研究挑战了“冻结世界假设”，并在两种现实动态场景下评估LRM的鲁棒性：中断测试（在预算约束输出下评估模型响应的准确性）与动态上下文测试（评估模型对运行中变化的适应能力）。在需要长程推理的数学与编程基准测试中，静态评估结果始终高估鲁棒性：即使是在静态设置中达到高准确率的最先进LRM，当遭遇中断或上下文变化时也可能不可预测地失败，尤其是在推理后期引入更新信息时，性能下降幅度可达60%。我们的分析进一步揭示了多种新型失效模式，包括推理泄露（模型将被中断的推理过程折叠入最终答案）、恐慌（时间压力下模型完全放弃推理并返回错误答案），以及自我怀疑（尝试整合更新信息时性能恶化）。项目页面：http://dynamic-lm.github.io/