Report
Google Cloud AI Research releases Physical-Engineering’s Last Exam
Google Cloud AI Research says all frontier models scored under 50% on its physical-engineering test.
TLDR
Google Cloud AI Research announced Physical-Engineering’s Last Exam (PLE), which it says tests work that builds the physical world. It says all frontier models scored under 50% and contrasts PLE with HLE’s knowledge tests and ALE’s digital-work tests.
Combined views
2.5K
4 Sources, first seen ago
