AI struggles to reliably reproduce scientific experiments across fields
PAPERBENCHX: Can AI Reproduce Science Across Disciplines?
Artificial Intelligence
Summary
Before AI can be trusted to make new scientific discoveries, it should first be able to recreate experiments that scientists have already done. The authors created PAPERBENCHX, a test with 93 tasks from 12 different science areas like biology and chemistry. AI agents tried to reproduce these experiments using specialized scientific software, but the best AI only fully succeeded 14% of the time. This shows that while AI can understand scientific papers to some extent, it still has a hard time producing trustworthy scientific results.
What this means in practice
- •For scientific software developers: Create better AI tools that can operate specialized scientific software for automating reproducible experiments across scientific domains.
- •For science workflow engineers: Design workflows that support diagnosing and validating long-running scientific computations driven by AI agents.
Authors
Ruoyu Wu, Zengji Tu, Aolong Sun, Tianyi Ma, Junren Li, Xinbo Xu, Li Fu, Ziyan Zhang, Pinrui Huang, Xiaochen Wang, Chengda Wu, Yutao Zhou, Jianjun Wu, Lin Chang, Baobao Chang, Liang Chen, Kuan Li
Abstract
Before AI can be trusted to discover new science, it should be able to reproduce what science already knows. Reproduction provides a rigorous checkpoint for an AI scientist: the target claim is known, yet recovering it still requires translating an incomplete scientific description into an executable experiment and producing evidence that can be independently verified. Existing research-reproduction benchmarks largely focus on AI and machine learning, where experiments share relatively homogeneous software stacks. Extending this capability across scientific disciplines introduces a different challenge: agents must operate domain-specific scientific software, make physical and numerical choices, diagnose long-running computations, and distinguish valid evidence from merely plausible outputs. We introduce PAPERBENCHX, a benchmark of 93 expert-constructed paper-reproduction tasks spanning 12 research areas across electromagnetics, photonics, chemistry, materials science, biology, and robotics. Each task provides the source paper, a public task specification, and a domain-native scientific environment involving specialized scientific software. Agents must reconstruct a key experiment, execute and diagnose it, and deliver a runnable workflow that supports the target scientific claim. To evaluate scientific validity rather than surface-level agreement, we replay every submission from scratch and score only regenerated evidence against expert-validated criteria. Across 10 frontier agent configurations, the best configuration fully reproduces only 14% tasks under the 90% threshold. Meanwhile, mean rubric scores range from 27.53% to 59.08%, indicating that substantial partial progress rarely translates into complete scientific reproduction. These results reveal a large gap between paper understanding and reliably producing reproducible scientific evidence.