EvoArena: Tracking Memory Evolution for Robust LLM Agents in Dynamic Environments

2026-06-11

AI Summary

Existing large language model agents struggle to adapt to dynamically evolving real-world environments. To address this limitation, this work proposes EvoArena—a dynamic evaluation benchmark that simulates the gradual evolution of terminals, software, and societal preferences—and introduces a novel formulation of environmental change as a sequence of continuous updates. Furthermore, the authors design EvoMem, a memory mechanism that employs structured logging of memory update histories and patch-based storage to enable agents to reason about environmental shifts and support continual learning. Experimental results demonstrate that EvoMem improves average accuracy by 1.5% on EvoArena, with gains of 6.1% and 4.8% on GAIA and LoCoMo benchmarks, respectively, and achieves a 3.7% improvement in chain-level task accuracy, significantly enhancing agents’ adaptability to dynamic environments.

Abstract

Large language model (LLM) agents have achieved strong performance on a wide range of benchmarks, yet most evaluations assume static environments. In contrast, real-world deployment is inherently dynamic, requiring agents to continually align their knowledge, skills, and behavior with changing environments and updated task conditions. To address this gap, we introduce EvoArena, a benchmark suite that models environment changes as sequences of progressive updates across terminal, software, and social domains. We further propose EvoMem, a patch-based memory paradigm that records memory evolution as structured update histories, enabling agents to reason about environmental evolution through changes in their memory. Experiments show that current agents struggle on EvoArena, achieving an average accuracy of 39.6% across evolving terminal, software, and social-preference domains. EvoMem consistently improves performance, yielding an average gain of 1.5% on EvoArena and also improving standard benchmarks such as GAIA and LoCoMo by 6.1% and 4.8%. Beyond individual tasks, EvoMem further improves chain-level accuracy by 3.7% on EvoArena, where success requires completing a consecutive sequence of related evolutionary subtasks. Mechanistic analysis shows that EvoMem improves evidence capture in the memory, indicating better preservation of complete evolving environment states. Our results highlight the importance of modeling evolution in both evaluation and memory for reliable agent deployment.

Problem

• dynamic environments
• LLM agents
• memory evolution
• environmental change
• robustness

Innovation

• EvoArena
• EvoMem
• memory evolution
• dynamic environments
• LLM agents