OpenMAS-GCom is introduced, a benchmark for diagnosing how these components affect G-MAS performance through controlled interventions, and 400 G-MAS-Complex tasks requiring agents to combine information from multiple documents, resolve conflicting records, and return specified values with source identifiers are added.
Abstract
Graph-enhanced multi-agent systems (G-MAS) coordinate large language model agents through communication graphs and role assignments, which determine how agents exchange information and divide responsibilities. However, final-score comparisons across systems combine differences in models, communication patterns, roles, and computation costs, making performance differences difficult to attribute to specific communication structures, role assignments, and information flows. To address this evaluation attribution problem, we introduce OpenMAS-GCom, a benchmark for diagnosing how these components affect G-MAS performance through controlled interventions. We represent systems through collaboration units, communication links, shared intermediate information, and execution rules. OpenMAS-GCom compares original systems with versions modified by changing one component while keeping tasks, models, prompts, and budget limits fixed. We rewire communication edges, remove specialist or critic agents, replace intermediate messages with incorrect content, and disable workers during execution. The benchmark evaluates 17 single-agent, ordinary multi-agent, and graph-enhanced configurations on 29 datasets across six domains. We add 400 G-MAS-Complex tasks requiring agents to combine information from multiple documents, resolve conflicting records, and return specified values with source identifiers. Experiments show larger mean losses after specialist removal than after critic removal, different performance degradation under incorrect messages and worker failures despite similar original scores, and different configurations achieving the highest accuracy and accuracy per token on G-MAS-Complex.
This work introduces a generalizable evaluation framework that maps native MAS traces into a shared space of unified collaboration graphs, enabling different methods to be evaluated under the same representation, reference set, and metric panel.
LLM-based multi-agent systems coordinate specialized reasoning through aggregation, interaction, and adaptive control, yet their potential for graph learning remains unexplored. Graph learning is a natural setting for such systems because useful evidence may arise from heterogeneous local, long-range, global structural...
Jia-Yi Yang, Yi-Fang Chen, Yuan-Fu Sun et al.· arXiv.org· 0 citations
MACE Loop is a memory-agent co-evolution framework that adapts memory organization and agent memory use through execution feedback that outperforms ten baselines with an average score of 81.11%, compared with 78.97% for the strongest baseline, SAGE.
Kai-Rui Yang, Ming-Hao An, Xun-Kai Li et al.· 0 citations
This work synthesizes a task-conditioned temporal workflow graph that jointly specifies agent connectivity and edge-level communication semantics, and introduces ReActNet, a training-free framework that compiles a query and a set of role-specialized agents into a sequence of directed communication graphs.
E2-Explainer is proposed, a model-agnostic framework for providing interpretable explanations of communication topologies produced by arbitrary topology generators that identifies compact communication subgraphs supported by edge-level evidence of task preservation.
This work introduces a controlled and MAS-demanding diagnostic benchmark for representative MAS efficiency methods and shows that many reported gains are setup-dependent and may arise from structural collapse, disabled tool pathways, or starting systems where random pruning already preserves accuracy, rather than robus...
Jia-Mu Zhang, Ling-Xi Zhang, Peng-Jun Lu et al.· 2 citations
Related blog posts
Microsoft Research Blog· microsoft.comNov 16, 2026
AI coding agents are increasingly help developers fix bugs and build features largely on their own, but every interaction with their models consumes tokens. In each session, an agent sends instructions, tool descriptions, and session history to its model. Most efforts to reduce this cost focus on individual tasks. Yet developers return to the same […] The post TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient appeared first on Microsoft Research.
MIT News · Artificial Intelligence· news.mit.eduNov 13, 2026