Scaling Teams or Scaling Time? Memory Enabled Lifelong Learning in LLM Multi-Agent Systems

TL;DR

This paper introduces LLMA-Mem, a memory framework for LLM multi-agent systems, demonstrating how memory design influences the trade-offs between team size and lifelong learning for better long-term performance.

Contribution

It presents a novel lifelong memory framework and analyzes the non-monotonic effects of team size and memory design on system performance under cost constraints.

Findings

LLMA-Mem improves long-horizon performance across tasks.

Memory design influences the effectiveness of team scaling.

Smaller teams with better memory can outperform larger teams.

Abstract

Large language model (LLM) multi-agent systems can scale along two distinct dimensions: by increasing the number of agents and by improving through accumulated experience over time. Although prior work has studied these dimensions separately, their interaction under realistic cost constraints remains unclear. In this paper, we introduce a conceptual scaling view of multi-agent systems that jointly considers team size and lifelong learning ability, and we study how memory design shares this landscape. To this end, we propose \textbf{LLMA-Mem}, a lifelong memory framework for LLM multi-agent systems under flexible memory topologies. We evaluate LLMA-Mem on \textsc{MultiAgentBench} across coding, research, and database environments. Empirically, LLMA-Mem consistently improves long-horizon performance over baselines while reducing cost. Our analysis further reveals a non-monotonic scaling landscape: larger teams do not always produce better long-term performance, and smaller teams can outperform larger ones when memory better supports the reuse of experience. These findings position memory design as a practical path for scaling multi-agent systems more effectively and more efficiently over time.