OrchMAS: Orchestrated Reasoning with Multi Collaborative Heterogeneous Scientific Expert Structured Agents

TL;DR

OrchMAS introduces a dynamic, multi-tier multi-model orchestration framework that enhances scientific reasoning by enabling flexible, specialized agent collaboration, iterative planning, and domain adaptation, outperforming existing multi-agent systems.

Contribution

This work presents a novel, domain-oriented, interactive multi-model orchestration framework that dynamically constructs reasoning pipelines and adapts roles for scientific tasks, improving flexibility and robustness.

Findings

Significant performance improvements over existing systems.

Effective dynamic replanning and role reallocation.

Enhanced robustness in scientific reasoning tasks.

Abstract

Multi-agent large language model frameworks are promising for complex multi step reasoning, yet existing systems remain weak for scientific and knowledge intensive domains due to static prompts and agent roles, rigid workflows, and homogeneous model reliance, leading to poor domain adaptation, limited reasoning flexibility, and high latency on heterogeneous or long-horizon scientific tasks. They also struggle to revise earlier decisions when intermediate reasoning diverges, reducing reliability in structured and calculation heavy settings. To address these limitations, we propose a scientific domain oriented interactive two tier multi model orchestration framework. A dedicated orchestration model analyzes each task, dynamically constructs a domain aware reasoning pipeline, and instantiates specialized expert agents with tailored prompts, while an execution model performs each step under generated role and instruction specifications. The orchestrator iteratively updates the pipeline based on intermediate feedback, enabling dynamic replanning, role reallocation, and prompt refinement across multi turn interactions, strengthening robustness and specialization for scientific reasoning through structured heterogeneous model collaboration. The framework is model agnostic and supports heterogeneous LLM integration with different capacities or costs, enabling flexible performance efficiency trade offs in practical scientific deployments. Experiments show consistent improvements over existing multi agent systems and strong baselines across diverse reasoning and scientific style benchmarks.