Crossmark

Systematic benchmarking demonstrates large language models have not reached the diagnostic accuracy of traditional rare-disease decision support tools

Crossref DOI link: https://doi.org/10.1038/s41431-026-02054-5

Published Online: 2026-02-24

Published Print: 2026-04

Update policy: https://doi.org/10.1007/springer_crossmark_policy

Authors

Reese, Justin T. https://orcid.org/0000-0002-2170-2250
Chimirri, Leonardo https://orcid.org/0000-0002-6912-8518
Bridges, Yasemin

Danis, Daniel

Caufield, J. Harry https://orcid.org/0000-0001-5705-7831
Gargano, Michael A.

Kroll, Carlo

Schmeder, Andrew

Liu, Fengchen

Wissink, Kyran

McMurry, Julie A. https://orcid.org/0000-0002-9353-5498
Graefe, Adam S. L.

Niyonkuru, Enock

Korn, Daniel R.

Casiraghi, Elena https://orcid.org/0000-0003-2024-7572
Valentini, Giorgio https://orcid.org/0000-0002-5694-3919
Jacobsen, Julius O. B.

Haendel, Melissa https://orcid.org/0000-0001-9114-8737
Smedley, Damian https://orcid.org/0000-0002-5836-9850
Mungall, Christopher J.

Robinson, Peter N. https://orcid.org/0000-0002-0736-9199
Funding

Funding for this research was provided by:

U.S. Department of Health & Human Services | NIH | Eunice Kennedy Shriver National Institute of Child Health and Human Development (5R01HD103805)

U.S. Department of Health & Human Services | NIH | National Human Genome Research Institute (5R24OD011883)
License Information

Text and Data Mining valid from 2026-02-24

Version of Record valid from 2026-02-24
More Information

Article History

Received: 30 June 2025

Revised: 15 January 2026

Accepted: 9 February 2026

First Online: 24 February 2026

Competing interests

: The authors declare no competing interests.

: Ethical approval was not required as only published data was analyzed. This study followed Consolidated Reporting Guidelines for Prognostic and Diagnostic Machine Learning Modeling Studies [ ].

Document is current

Any future updates will be listed below