Toward a test of medical AI superintelligence

Researchers led by E. Goh and colleagues at Google, Meta, Amazon, Anthropic, OpenAI, and Microsoft argue in Nature Medicine that existing benchmarks for medical AI are misleading and insufficient, urging the development of a rigorous, task-based framework to define and measure medical AI superintelligence.

Researchers urgently need a rigorous, task-based framework to define and measure medical AI ‘superintelligence’, because existing benchmarks are misleading and insufficient. This is a preview of subscription content, access via your institution https://wayf.springernature.com?redirect uri=https%3A%2F%2Fwww.nature.com%2Farticles%2Fs41591-026-04539-8 Access options Access Nature and 54 other Nature Portfolio journals Get Nature+, our best-value online-access subscription 27,99 € / 30 days cancel any time Subscribe to this journal Receive 12 print issues and online access 269,00 € per year only 22,42 € per issue Buy this article - Purchase on SpringerLink - Instant access to the full article PDF. 39,95 € Prices may be subject to local taxes which are calculated during checkout References American Medical Association. AMA https://go.nature.com/4ypndAg https://go.nature.com/4ypndAg 12 March 2026 .Nori, H. et al. Preprint at https://doi.org/10.48550/arxiv.2506.22405 https://doi.org/10.48550/arxiv.2506.22405 2025 .Palmer, K. STAT https://go.nature.com/4vHS5cB https://go.nature.com/4vHS5cB 12 January 2026 .Goh, E. et al. JAMA Netw. Open 7 , e2440969 2024 .Goh, E. et al. Nat. Med. 31 , 1233–1238 2025 .Brodeur, P. G. et al. Science 392 , 524–527 2026 .Sellen, A. & Horvitz, E. Commun. ACM 67 , 18–23 2024 .Adler-Milstein, J. et al. JAMA 331 , 1173 2024 .Hayat, H. et al. Preprint at medRxiv https://doi.org/10.1101/2025.07.14.25331406 https://doi.org/10.1101/2025.07.14.25331406 2025 .Ramaswamy, A. et al. Nat. Med. 32 , 1671–1675 2026 .Navarro, D. F. et al. Preprint at https://doi.org/10.48550/arxiv.2603.11413 https://doi.org/10.48550/arxiv.2603.11413 2026 .Wu, D. J. et al. BMJ Digital Health AI 2 , e000032 2026 .Jiang, Y. et al. NEJM AI https://doi.org/10.1056/AIdbp2500144 https://doi.org/10.1056/AIdbp2500144 2025 .Bedi, S. et al. Nat. Med. 32 , 943–951 2026 .Gommers, J. et al. Lancet 407 , 505–514 2026 . Author information Authors and Affiliations Corresponding author Ethics declarations Competing interests Y.L., P.-H.C.C. and M.S. are employees of Google. J.R. is an employee of Meta. D.L.-M. is an employee of Amazon. P.M.-Y. is an employee of Anthropic. K.S. is an employee of OpenAI. E.H. is an employee of Microsoft. E.G. and A.R. report consulting fees from Google. L.G.M. and D.W. report consulting fees from Meta. J.H.C. reports co-founding Reaction Explorer, which develops and licenses organic chemistry education software; and receiving consulting fees from Sutton Pierce, Younker Hyde Macfarlane and Sykes McAllister as a medical expert witness. J.H.C. receives funding from the National Institutes of Health NIH /National Institute of Allergy and Infectious Diseases NIAID , NIH/National Center for Advancing Translational Services NCATS Clinical and Translational Science Award, NIH/Center for Undiagnosed Diseases at Stanford, Stanford Bio-X, Stanford RAISE Health, the Josiah Macy Jr Foundation and the Stanford CARE AI Scholar Fellowship. The other authors declare no competing interests. Supplementary information Supplementary Information download PDF https://media.springernature.com/original/springer-static/esm/art%3A10.1038%2Fs41591-026-04539-8/MediaObjects/41591 2026 4539 MOESM1 ESM.pdf Supplementary Tables 1 and 2 Rights and permissions About this article Cite this article Goh, E., Wu, D., Walton, C. et al. Toward a test of medical AI superintelligence. Nat Med 2026 . https://doi.org/10.1038/s41591-026-04539-8 Published: Version of record: DOI: https://doi.org/10.1038/s41591-026-04539-8