Memory for Autonomous LLM Agents:Mechanisms, Evaluation, and Emerging Frontiers

TL;DR

This survey reviews how memory mechanisms are designed, evaluated, and applied in modern LLM agents, highlighting recent advances, evaluation shifts, and future challenges in creating adaptive, persistent AI systems.

Contribution

It provides a formal framework and taxonomy for understanding memory in LLM agents, and analyzes recent mechanisms, benchmarks, and applications in this rapidly evolving field.

Findings

Shift from static to multi-session memory benchmarks

Identification of five key memory mechanism families

Highlighting open challenges like continual learning and multimodal memory

Abstract

Large language model (LLM) agents increasingly operate in settings where a single context window is far too small to capture what has happened, what was learned, and what should not be repeated. Memory -- the ability to persist, organize, and selectively recall information across interactions -- is what turns a stateless text generator into a genuinely adaptive agent. This survey offers a structured account of how memory is designed, implemented, and evaluated in modern LLM-based agents, covering work from 2022 through early 2026. We formalize agent memory as a \emph{write--manage--read} loop tightly coupled with perception and action, then introduce a three-dimensional taxonomy spanning temporal scope, representational substrate, and control policy. Five mechanism families are examined in depth: context-resident compression, retrieval-augmented stores, reflective self-improvement, hierarchical virtual context, and policy-learned management. On the evaluation side, we trace the shift from static recall benchmarks to multi-session agentic tests that interleave memory with decision-making, analyzing four recent benchmarks that expose stubborn gaps in current systems. We also survey applications where memory is the differentiating factor -- personal assistants, coding agents, open-world games, scientific reasoning, and multi-agent teamwork -- and address the engineering realities of write-path filtering, contradiction handling, latency budgets, and privacy governance. The paper closes with open challenges: continual consolidation, causally grounded retrieval, trustworthy reflection, learned forgetting, and multimodal embodied memory.