MAS-Orchestra: Understanding and Improving Multi-Agent Reasoning Through Holistic Orchestration and Controlled Benchmarks

TL;DR

This paper introduces MAS-Orchestra, a reinforcement learning framework for holistic multi-agent system orchestration, and MASBENCH, a benchmark for analyzing when multi-agent systems outperform single-agent systems.

Contribution

It presents a novel training framework for global MAS orchestration and a controlled benchmark to study MAS benefits based on task structure and capabilities.

Findings

MAS-Orchestra improves performance on reasoning and QA tasks.

MAS gains depend on task structure and verification protocols.

Over 10x efficiency gains over baselines.

Abstract

While multi-agent systems (MAS) promise elevated intelligence through coordination of agents, current approaches to automatic MAS design under-deliver. Such shortcomings stem from two key factors: (1) methodological complexity - agent orchestration is performed using sequential, code-level execution that limits global system-level holistic reasoning and scales poorly with agent complexity - and (2) efficacy uncertainty - MAS are deployed without understanding if there are tangible benefits compared to single-agent systems (SAS). We propose MASOrchestra, a training-time framework that formulates MAS orchestration as a function-calling reinforcement learning problem with holistic orchestration, generating an entire MAS at once. In MAS-Orchestra, complex, goal-oriented subagents are abstracted as callable functions, enabling global reasoning over system structure while hiding internal execution details. To rigorously study when and why MAS are beneficial, we introduce MASBENCH, a controlled benchmark that characterizes tasks along five axes: Depth, Horizon, Breadth, Parallel, and Robustness. Our analysis reveals that MAS gains depend critically on task structure, verification protocols, and the capabilities of both orchestrator and subagents, rather than holding universally. Guided by these insights, MAS-Orchestra achieves consistent improvements on public benchmarks including mathematical reasoning, multi-hop QA, and search-based QA, while achieving more than 10x efficiency over strong baselines. Together, MAS-Orchestra and MASBENCH enable better training and understanding of MAS in the pursuit of multi-agent intelligence.