EvoMaster: A Foundational Evolving Agent Framework for Agentic Science at Scale

TL;DR

EvoMaster is a scalable, self-evolving agent framework designed to facilitate autonomous scientific discovery across multiple disciplines, demonstrating state-of-the-art performance on key benchmarks.

Contribution

The paper introduces EvoMaster, a domain-agnostic framework enabling agents to iteratively refine hypotheses and knowledge, significantly advancing agentic science at scale.

Findings

Achieves state-of-the-art scores on four scientific benchmarks.

Outperforms baseline OpenClaw with 159%-316% improvements.

Easily scalable, requiring approximately 100 lines of code.

Abstract

The convergence of large language models and agents is catalyzing a new era of scientific discovery: Agentic Science. While the scientific method is inherently iterative, existing agent frameworks are predominantly static, narrowly scoped, and lack the capacity to learn from trial and error. To bridge this gap, we present EvoMaster, a foundational evolving agent framework engineered specifically for Agentic Science at Scale. Driven by the core principle of continuous self-evolution, EvoMaster empowers agents to iteratively refine hypotheses, self-critique, and progressively accumulate knowledge across experimental cycles, faithfully mirroring human scientific inquiry. Crucially, as a domain-agnostic base harness, EvoMaster is exceptionally easy to scale up -- enabling developers to build and deploy highly capable, self-evolving scientific agents for arbitrary disciplines in approximately 100 lines of code. Built upon EvoMaster, we incubated the SciMaster ecosystem across domains such as machine learning, physics, and general science. Evaluations on four authoritative benchmarks (Humanity's Last Exam, MLE-Bench Lite, BrowseComp, and FrontierScience) demonstrate that EvoMaster achieves state-of-the-art scores of 41.1%, 75.8%, 73.3%, and 53.3%, respectively. It comprehensively outperforms the general-purpose baseline OpenClaw with relative improvements ranging from +159% to +316%, robustly validating its efficacy and generality as the premier foundational framework for the next generation of autonomous scientific discovery. EvoMaster is available at https://github.com/sjtu-sai-agents/EvoMaster.