Zhipu AI
GLM-5
9 benchmarks
AIME
#14 of 39
95.83%
AIME
#14 of 39
92.7%
GPQA Diamond
#22 of 49
86.0%
SWE-bench Verified
#11 of 40
77.8%
BrowseComp
#10 of 17
75.9%
Terminal-Bench 2.0
#11 of 14
56.2%
Humanity's Last Exam
#7 of 24
50.4%
Humanity's Last Exam
#7 of 24
30.5%
MRCR v2
#5 of 5
26.3%
benchmark.space canonical public ledger of AI capability