https://policywindow.org/wiki/frontiermath Policy Window Editorial Board Policy Window Policy Window 2024 Reference Work AI governance AI policy regulation evidence catalog 2024-01-01 2026-05-30 en https://policywindow.org/wiki/frontiermath frontiermath https://policywindow.org/wiki/frontiermath?asOf=2024-01-01 https://policywindow.org/wiki/methodology text/html application/json application/ld+json Creative Commons Attribution 4.0 International (CC BY 4.0) FrontierMath — capability evaluation benchmark. Hundreds of original research-mathematician-curated math problems requiring deep reasoning. Held-out evaluation only.