Autonomous agents are increasingly expected to support scientific research, and recent benchmarks report progress in code repair and autonomous experimentation. However, these evaluations typically assume a pre-configured execution environment, which requires resolving complex software dependencies, aligning hardware and framework versions, and configuring distributed execution, yet this capability remains largely unbenchmarked. We introduce ResearchEnvBench, a benchmark for environment synthesis in research code execution. Given a research repository, documentation, and a target execution setting, agents must construct an environment that successfully executes at runtime. Evaluations on diverse research repositories reveal a substantial gap in current SOTA agents, with failures dominated by incomplete dependency resolution and brittle version coupling. ResearchEnvBench provides a realistic testbed for advancing autonomous agents toward reproducible scientific research.
@article{arxiv.2603.06739,
title = {ResearchEnvBench: Benchmarking Agents on Environment Synthesis for Research Code Execution},
author = {Yubang Wang and Chenxi Zhang and Bowen Chen and Zezheng Huai and Zihao Dai and Xinchi Chen and Yuxin Wang and Yining Zheng and Jingjing Gong and Xipeng Qiu},
journal= {arXiv preprint arXiv:2603.06739},
year = {2026}
}