Metraipolis › archive › dc_f371c5e665a7
PAPERBENCHX: Can AI Reproduce Science Across Disciplines? — arXiv:2610.39744
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.39744
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-31T06:09:26.752Z
- SHA-256 of the archived text
fd89c10a2a42ce211321ed9b078cab02d200e2801dd8a47cb3bf729ca158ec05- Stable URL of this page
https://metraipolis.com/documento/dc_f371c5e665a7
How to cite
PAPERBENCHX: Can AI Reproduce Science Across Disciplines? — arXiv:2610.39744 — arXiv (en), https://arxiv.org/abs/2610.39744 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f371c5e665a7 (Metraipolis archive dc_f371c5e665a7, sha256 fd89c10a2a42ce211321ed9b078cab02d200e2801dd8a47cb3bf729ca158ec05)
BibTeX
@misc{dcf371c5e665a7,
title = {PAPERBENCHX: Can AI Reproduce Science Across Disciplines? — arXiv:2610.39744},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.39744},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f371c5e665a7; sha256 fd89c10a2a42ce211321ed9b078cab02d200e2801dd8a47cb3bf729ca158ec05},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.39744. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
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. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.39744 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-31T06:07:08.538Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).