Google
Deep Research Max
2 benchmarks
BrowseComp
#13 of 45
85.9%
Humanity's Last Exam
#9 of 72
54.6%
benchmark.space canonical public ledger of AI capability