cs.AIMar 8, 2026

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

Authors: Pengfei Du

Organizations: Hong Kong Research Institute of Technology, HongKong, China

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

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 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.

Figures & tables

SystemYearMemory CategoryDistinguishing Feature
RAG [ Lewis et al., 2020 ]2020Non-parametric retrievalFirst to couple a seq2seq generator with a dense document retriever at NeurIPS 2020.
RETRO [ Borgeaud et al., 2022 ]2022Retrieval at scaleChunks retrieved from a 2-trillion-token corpus; 7.5B-parameter model rivals 175B Jurassic-1 on 10/16 benchmarks.
ReAct [ Yao et al., 2022 ]2022Trajectory tracesReasoning-and-acting traces double as short-horizon working memory; 34% absolute gain on ALFWorld.
Reflexion [ Shinn et al., 2023 ]2023Reflective episodicVerbal self-critiques stored as episodic memory; 91% pass@1 on HumanEval (vs. 80% GPT-4 baseline).
Generative Agents [ Park et al., 2023 ]2023Episodic + reflective25 simulated characters autonomously organize a Valentine’s party via observation–reflection–planning cycles.
Voyager [ Wang et al., 2023a ]2023Procedural skill library3.3 ×\times more unique items and 15.3 ×\times faster tech-tree progression than prior Minecraft agents.
Table 1: Representative memory systems and benchmarks for LLM agents (2020–2026).
BenchmarkYearMulti-sessionMulti-turnAgentic tasksForgettingMultimodal
LoCoMo2024✓✓––✓
MemBench2025–✓–––
MemoryAgentBench2025–✓–✓–
MemoryArena2026✓✓✓––
Table 2: Feature comparison of recent agent memory benchmarks.

Explore similar work

CardsList
May 18, 2026

EvoMemBench: Benchmarking Agent Memory from a Self-Evolving Perspective

Recent benchmarks for Large Language Model (LLM) agents mainly evaluate reasoning, planning, and execution. However, memory is also essential for agents, as it enables them to store, update, and retrieve information over time. This ability remains under-evaluated, largely because existing benchmarks do not provide a systematic way to assess memory mechanisms. In this paper, we study agent memory from a self-evolving perspective and introduce EvoMemBench, a unified benchmark organized along two axes: memory scope (in-episode vs. cross-episode) and memory content (knowledge-oriented vs. execution-oriented). We compare 15 representative memory methods with strong long-context baselines under a standardized protocol. Results show that current memory systems are still far from a general solution: long-context baselines remain highly competitive, memory helps most when the current context is insufficient or tasks are difficult, and no single memory form works consistently across all settings. Retrieval-based methods remain strong for knowledge-intensive settings, whereas procedural and long-term memory methods are more effective for execution-oriented tasks when their stored experience matches the task structure. We hope EvoMemBench facilitates future research on more effective memory systems for LLM-based agents. Our code is available at https://github.com/DSAIL-Memory/EvoMemBench.

Yuyao Wang, Zhongjian Zhang, Mo Chi +7

Hong Kong University of Science and Technology (Guangzhou) · Beijing University of Posts and Telecommunications · Beijing Institute of Technology +1

LLM Agent MemoryLLM Agent EvaluationAI Agent BenchmarksMemory Agent BenchmarksAgent Benchmarks
AskPDF
Jul 15, 2026

Memory as a Controlled Process: Learned Adaptive Memory Management for LLM Agents

Large Language Model (LLM) agents increasingly rely on external memory systems to accumulate experience across tasks. Yet nearly all existing approaches, from graph-structured memories to reflective insight stores, access memory through fixed, hand-designed heuristics. We argue that this static view of memory is a core bottleneck for agentic learning because optimal memory behavior is fundamentally context-dependent. The early stages of the tasks, benefit from minimal retrieval because memory is sparse; recurring goal types benefit from plan reuse rather than generic nearest-neighbor lookup; stuck agents benefit from re-retrieval with alternative queries; and across long task streams, the memory store itself must be consolidated and pruned to remain useful. We present Memory as a Controlled Process (MemCon), a framework that models memory operations as a Markov Decision Process and learns an online policy that adaptively decides when, what, and how much to retrieve, when to inject a distilled plan, and when to consolidate or forget. MemCon is backend-agnostic: it wraps any existing memory implementation, learns from task-by-task binary feedback with no pretraining and no additional LLM calls, and uses a lightweight tabular contextual bandit with UCB exploration that converges within tens of tasks. Across 6 benchmarks, 3 agent frameworks, and 3 LLM backbones, MemCon consistently outperforms multiple memory baselines by up to 15.2 points in task success while reducing token consumption by 5--20%.

Eric Hanchen Jiang, Zhi Zhang, Yuchen Wu +11

University of California Los Angeles · University of Washington · Northwestern University

Persistent Memory for Language ModelsAgent Memory ManagementMemory-Augmented Language Model Agents
AskPDF
Feb 22, 2026

Anatomy of Agentic Memory: Taxonomy and Empirical Analysis of Evaluation and System Limitations

Agentic memory systems enable large language model (LLM) agents to maintain state across long interactions, supporting long-horizon reasoning and personalization beyond fixed context windows. Despite rapid architectural development, the empirical foundations of these systems remain fragile: existing benchmarks are often underscaled, evaluation metrics are misaligned with semantic utility, performance varies significantly across backbone models, and system-level costs are frequently overlooked. This survey presents a structured analysis of agentic memory from both architectural and system perspectives. We first introduce a concise taxonomy of MAG systems based on four memory structures. Then, we analyze key pain points limiting current systems, including benchmark saturation effects, metric validity and judge sensitivity, backbone-dependent accuracy, and the latency and throughput overhead introduced by memory maintenance. By connecting the memory structure to empirical limitations, this survey clarifies why current agentic memory systems often underperform their theoretical promise and outlines directions for more reliable evaluation and scalable system design.

Dongming Jiang, Yi Li, Songtao Wei +8

University of Texas at Dallas · University of California, Davis · Texas A&M University

Agent MemoryAgent Memory ManagementAI Agent EvaluationMemory Agent Benchmarks
AskPDF