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

arXiv: https://arxiv.org/abs/2601.14652 · HTML: https://mlsanity.com/paper/2601.14652/

## Key facts

- Cited by 32 papers in the map (data through 14 Oct 2026).
- Builds on 29 of its 42 references that are in the map.
- First posted on arXiv 21 Jan 2026.

## Builds on (most cited first, 29 of 29)

- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Self-Consistency Improves Chain of Thought Reasoning in Language Models](https://mlsanity.com/paper/2203.11171/) (2022)
- [HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering](https://mlsanity.com/paper/1809.09600/) (2018)
- [Self-Refine: Iterative Refinement with Self-Feedback](https://mlsanity.com/paper/2303.17651/) (2023)
- [GPQA: A Graduate-Level Google-Proof Q&A Benchmark](https://mlsanity.com/paper/2311.12022/) (2023)
- [Improving Factuality and Reasoning in Language Models through Multiagent Debate](https://mlsanity.com/paper/2305.14325/) (2023)
- [The Rise and Potential of Large Language Model Based Agents: A Survey](https://mlsanity.com/paper/2309.07864/) (2023)
- [RULER: What's the Real Context Size of Your Long-Context Language Models?](https://mlsanity.com/paper/2404.06654/) (2024)
- [Large Language Models Cannot Self-Correct Reasoning Yet](https://mlsanity.com/paper/2310.01798/) (2023)
- [Automated Design of Agentic Systems](https://mlsanity.com/paper/2408.08435/) (2024)
- [A Dynamic LLM-Powered Agent Network for Task-Oriented Agent Collaboration](https://mlsanity.com/paper/2310.02170/) (2023)
- [Scaling Large Language Model-based Multi-Agent Collaboration](https://mlsanity.com/paper/2406.07155/) (2024)
- [AI Agents That Matter](https://mlsanity.com/paper/2407.01502/) (2024)
- [BrowseComp-Plus: A More Fair and Transparent Evaluation Benchmark of Deep-Research Agent](https://mlsanity.com/paper/2508.06600/) (2025)
- [More Agents Is All You Need](https://mlsanity.com/paper/2402.05120/) (2024)
- [Multi-agent Architecture Search via Agentic Supernet](https://mlsanity.com/paper/2502.04180/) (2025)
- [Multi-Agent Design: Optimizing Agents with Better Prompts and Topologies](https://mlsanity.com/paper/2502.02533/) (2025)
- [Multi-Agent Collaboration via Evolving Orchestration](https://mlsanity.com/paper/2505.19591/) (2025)
- [Router-R1: Teaching LLMs Multi-Round Routing and Aggregation via Reinforcement Learning](https://mlsanity.com/paper/2506.09033/) (2025)
- [Physics of Language Models: Part 2.2, How to Learn From Mistakes on Grade-School Math Problems](https://mlsanity.com/paper/2408.16293/) (2024)
- [MAS-GPT: Training LLMs to Build LLM-based Multi-Agent Systems](https://mlsanity.com/paper/2503.03686/) (2025)
- [Single-agent or Multi-agent Systems? Why Not Both?](https://mlsanity.com/paper/2505.18286/) (2025)
- [MAS-ZERO: Designing Multi-Agent Systems with Zero Supervision](https://mlsanity.com/paper/2505.14996/) (2025)
- [Weak-for-Strong: Training Weak Meta-Agent to Harness Strong Executors](https://mlsanity.com/paper/2504.04785/) (2025)
- [Controlling Performance and Budget of a Centralized Multi-agent LLM System with Reinforcement Learning](https://mlsanity.com/paper/2511.02755/) (2025)
- [How to Train a Leader: Hierarchical Reasoning in Multi-Agent LLMs](https://mlsanity.com/paper/2507.08960/) (2025)
- [The Era of Agentic Organization: Learning to Organize with Language Models](https://mlsanity.com/paper/2510.26658/) (2025)

## Built on by (most cited first, 30 of 32)

- [Natural-Language Agent Harnesses](https://mlsanity.com/paper/2603.25723/) (2026)
- [Single-Agent LLMs Outperform Multi-Agent Systems on Multi-Hop Reasoning Under Equal Thinking Token Budgets](https://mlsanity.com/paper/2604.02460/) (2026)
- [AgentArk: Distilling Multi-Agent Intelligence into a Single LLM Agent](https://mlsanity.com/paper/2602.03955/) (2026)
- [The Illusion of Multi-Agent Advantage](https://mlsanity.com/paper/2606.13003/) (2026)
- [MAS-ProVe: Understanding the Process Verification of Multi-Agent Systems](https://mlsanity.com/paper/2602.03053/) (2026)
- [Skill-MAS: Evolving Meta-Skill for Automatic Multi-Agent Systems](https://mlsanity.com/paper/2606.18837/) (2026)
- [Unified-MAS: Universally Generating Domain-Specific Nodes for Empowering Automatic Multi-Agent Systems](https://mlsanity.com/paper/2603.21475/) (2026)
- [Scaling Teams or Scaling Time? Memory Enabled Lifelong Learning in LLM Multi-Agent Systems](https://mlsanity.com/paper/2604.03295/) (2026)
- [EvoHarnessBench: Can Your Agents Keep Pace with an Evolving Harness?](https://mlsanity.com/paper/2609.04280/) (2026)
- [Experience as a Compass: Multi-agent RAG with Evolving Orchestration and Agent Prompts](https://mlsanity.com/paper/2604.00901/) (2026)
- [Graph Engineering in the Era of LLM Agents: From Individual Intelligence to System Intelligence](https://mlsanity.com/paper/2608.21156/) (2026)
- [Reward Modeling for Multi-Agent Orchestration](https://mlsanity.com/paper/2606.13598/) (2026)
- [Markets, Not Planners: Decentralized Orchestration of LLM Agents with Private Information](https://mlsanity.com/paper/2608.23867/) (2026)
- [Shared Prefixes, Better Credit: Adaptive Routing for Multi-Agent Reasoning](https://mlsanity.com/paper/2608.02291/) (2026)
- [Advancing MLLM-based UAV Image Understanding and Reasoning: A Benchmark and a Training-Free Multi-Agent System](https://mlsanity.com/paper/2608.11738/) (2026)
- [SwarmBench: Can Large Language Models Act as Agent Swarm Orchestrators?](https://mlsanity.com/paper/2608.30661/) (2026)
- [CoMemBench: Benchmarking Collaborative Memory Boundaries across Multi-Agent Workflow Topologies](https://mlsanity.com/paper/2609.32192/) (2026)
- [Self-Retrospection Distillation: Turning Post-hoc Experiences into Prior Foresight](https://mlsanity.com/paper/2610.08077/) (2026)
- [OrchBench: Evaluating Multi-Agent Orchestration Plans in Isolation via Deterministic Simulation](https://mlsanity.com/paper/2607.25656/) (2026)
- [EASy: Towards Efficient LLM-Based Agentic System](https://mlsanity.com/paper/2608.04588/) (2026)
- [Enhancing Human Mobility Prediction with Spatially Aware LLM-based Multi-Agent Systems](https://mlsanity.com/paper/2609.14227/) (2026)
- [CERA-MoA: Co-Evolving Routing Mechanisms with Continually Learning LLM Agents](https://mlsanity.com/paper/2609.18779/) (2026)
- [Multi-Agent System Search via Active Substructure-aware Policy Optimization](https://mlsanity.com/paper/2609.32430/) (2026)
- [CAIRN: Dynamic Fact-Intent DAGs for Multi-Agent Exploration](https://mlsanity.com/paper/2609.32700/) (2026)
- [Visual Parallel Search: Learning to Search High-Resolution Images with Parallel Tile Inspection and Adaptive Zoom](https://mlsanity.com/paper/2609.37002/) (2026)
- [Topological Coherence for Self-evolving Multi-agent Systems](https://mlsanity.com/paper/2609.37953/) (2026)
- [It Takes Workflows to Evolve Better Workflows](https://mlsanity.com/paper/2610.01026/) (2026)
- [Decoupled Multi-Agent Orchestration](https://mlsanity.com/paper/2610.07556/) (2026)
- [Proactive Agentic Orchestration](https://mlsanity.com/paper/2610.13485/) (2026)
- [MAS-Gate: Unlocking Multi-Agent System Lifecycle via Node-Centric Quantification](https://mlsanity.com/paper/2610.17613/) (2026)

## Abstract (arXiv)

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.