The FM Agent
Abstract
Large language models (LLMs) are catalyzing the development of autonomous AI research agents for scientific and engineering discovery. We present FM Agent, a novel and general-purpose multi-agent framework that leverages a synergistic combination of LLM-based reasoning and large-scale evolutionary search to address complex real-world challenges. The core of FM Agent integrates several key innovations: 1) a cold-start initialization phase incorporating expert guidance, 2) a novel evolutionary sampling strategy for iterative optimization, 3) domain-specific evaluators that combine correctness, effectiveness, and LLM-supervised feedback, and 4) a distributed, asynchronous execution infrastructure built on Ray. Demonstrating broad applicability, our system has been evaluated across diverse domains, including operations research, machine learning, GPU kernel optimization, and classical mathematical problems. FM Agent reaches state-of-the-art results autonomously, without human interpretation or tuning -- 1976.3 on ALE-Bench (+5.2\%), 43.56\% on MLE-Bench (+4.0pp), up to 20x speedups on KernelBench, and establishes new state-of-the-art(SOTA) results on several classical mathematical problems. Beyond academic benchmarks, FM Agent shows considerable promise for both large-scale enterprise R\&D workflows and fundamental scientific research, where it can accelerate innovation, automate complex discovery processes, and deliver substantial engineering and scientific advances with broader societal impact.
Cite
@article{arxiv.2510.26144,
title = {The FM Agent},
author = {Annan Li and Chufan Wu and Zengle Ge and Yee Hin Chong and Zhinan Hou and Lizhe Cao and Cheng Ju and Jianmin Wu and Huaiming Li and Haobo Zhang and Shenghao Feng and Mo Zhao and Fengzhi Qiu and Rui Yang and Mengmeng Zhang and Wenyi Zhu and Yingying Sun and Quan Sun and Shunhao Yan and Danyu Liu and Dawei Yin and Dou Shen},
journal= {arXiv preprint arXiv:2510.26144},
year = {2026}
}