• 🎉 Updates (2025-5-15) MasRouter is accpected to ACL 2025 Main!
• 🚩 Updates (2025-2-16) Initial upload to arXiv PDF.

MasRouter expands LLM routing to the multi-agent systems (MAS) for the first time. It leverages the powerful reasoning capabilities of LLM MAS, while also making it relatively cost-effective.

MasRouter integrates all components of MAS into a unified routing framework. It employs collaboration mode determination, role allocation, and LLM routing through a cascaded controller network, progressively constructing a MAS that balances effectiveness and efficiency.

Please download the GSM8K, HumanEval, MATH, MBPP, MMLU datasets and place it in the Datasets folder. The file structure should be organized as follows:

Datasets
└── gsm8k
    └── gsm8k.jsonl
└── humaneval
    └── humaneval-py.jsonl
└── MATH
    └── test
    └── train
└── mbpp
    └── mbpp.jsonl
└── MMLU
    └── data

Add API keys in template.env and change its name to .env. We recommend that this API be able to access multiple LLMs.

URL = "" # the URL of LLM backend
KEY = "" # the key for API

The code below verifies the experimental results of the mbpp dataset.

python experiments/run_mbpp.py

If you find this repo useful, please consider citing our paper as follows:

@misc{yue2025masrouter,
      title={MasRouter: Learning to Route LLMs for Multi-Agent Systems}, 
      author={Yanwei Yue and Guibin Zhang and Boyang Liu and Guancheng Wan and Kun Wang and Dawei Cheng and Yiyan Qi},
      year={2025},
      eprint={2502.11133},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2502.11133}, 
}

Special thanks to the following repositories for their invaluable code and datasets:

• MapCoder
• GPTSwarm.