Cut the Crap: An Economical Communication Pipeline for LLM-based Multi-Agent Systems

Recent advancements in large language model (LLM)-powered agents have shown that collective intelligence can significantly outperform individual capabilities, largely attributed to the meticulously designed inter-agent communication topologies. Though impressive in performance, existing multi-agent pipelines inherently introduce substantial token overhead, as well as increased economic costs, which pose challenges for their large-scale deployments. In response to this challenge, we propose an economical, simple, and robust multi-agent communication framework, termed $\mathtt{\text{AgentPrune}}$, which can seamlessly integrate into mainstream multi-agent systems and prunes redundant or even malicious communication messages. Technically, $\mathtt{\text{AgentPrune}}$ is the first to identify and formally define the $\mathit{\text{communication}}\mathit{\text{redundancy}}$ issue present in current LLM-based multi-agent pipelines, and efficiently performs one-shot pruning on the spatial-temporal message-passing graph, yielding a token-economic and high-performing communication topology. Extensive experiments across six benchmarks demonstrate that $\mathtt{\text{AgentPrune}}$ $\mathbf{\text{(I)}}$ achieves comparable results as state-of-the-art topologies at merely $\mathrm{\$}5.6$ cost compared to their $\mathrm{\$}43.7$, $\mathbf{\text{(II)}}$ integrates seamlessly into existing multi-agent frameworks with $28.1\mathrm{\%}\sim 72.8\mathrm{\%}\downarrow$ token reduction, and $\mathbf{\text{(III)}}$ successfully defend against two types of agent-based adversarial attacks with $3.5\mathrm{\%}\sim 10.8\mathrm{\%}\uparrow$ performance boost.