AgentSlimming: Towards Efficient and Cost-Aware Multi-Agent Systems

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Chen, Yulang, Peng, Haoxuan, Liu, Jinyan, Wen, Zichen, Liu, Dongrui, Zhang, Linfeng
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866914547793657856
author Chen, Yulang
Peng, Haoxuan
Liu, Jinyan
Wen, Zichen
Liu, Dongrui
Zhang, Linfeng
author_facet Chen, Yulang
Peng, Haoxuan
Liu, Jinyan
Wen, Zichen
Liu, Dongrui
Zhang, Linfeng
contents Large Language Model-based Multi-Agent Systems (MAS) have demonstrated remarkable capabilities in complex tasks. However, manually designing optimal communication topologies is labor-intensive, while automated expansion methods often result in bloated structures with redundant agents, leading to excessive token consumption. To address this problem, we introduce \textbf{AgentSlimming}, a plug-and-play compression framework for graph-structured multi-agent workflows. Motivated by pruning and quantization in neural networks, AgentSlimming compresses workflows by first estimating the importance score of each agent with a hybrid mechanism, and then removes redundant agents or replaces them with low-cost ones, where each operation is validated using a baseline-anchored acceptance rule to prevent performance collapse. Experiments show that AgentSlimming reduces average token cost by up to 78.9\% with negligible performance degradation, and sometimes even improves accuracy, achieving a strong Pareto-optimal trade-off between cost and quality. \textit{Our code is publicly available at https://github.com/CitrusYL/AgentSlimming
format Preprint
id arxiv_https___arxiv_org_abs_2605_08813
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle AgentSlimming: Towards Efficient and Cost-Aware Multi-Agent Systems
Chen, Yulang
Peng, Haoxuan
Liu, Jinyan
Wen, Zichen
Liu, Dongrui
Zhang, Linfeng
Machine Learning
Large Language Model-based Multi-Agent Systems (MAS) have demonstrated remarkable capabilities in complex tasks. However, manually designing optimal communication topologies is labor-intensive, while automated expansion methods often result in bloated structures with redundant agents, leading to excessive token consumption. To address this problem, we introduce \textbf{AgentSlimming}, a plug-and-play compression framework for graph-structured multi-agent workflows. Motivated by pruning and quantization in neural networks, AgentSlimming compresses workflows by first estimating the importance score of each agent with a hybrid mechanism, and then removes redundant agents or replaces them with low-cost ones, where each operation is validated using a baseline-anchored acceptance rule to prevent performance collapse. Experiments show that AgentSlimming reduces average token cost by up to 78.9\% with negligible performance degradation, and sometimes even improves accuracy, achieving a strong Pareto-optimal trade-off between cost and quality. \textit{Our code is publicly available at https://github.com/CitrusYL/AgentSlimming
title AgentSlimming: Towards Efficient and Cost-Aware Multi-Agent Systems
topic Machine Learning
url https://arxiv.org/abs/2605.08813