Tencent
Hunyuan Hy3
6 benchmarks
GPQA Diamond
#28 of 118
90.4%
SWE-bench Verified
#18 of 86
78.0%
MRCR v2
#7 of 18
75.1%
SWE-bench Pro
#17 of 43
57.9%
Humanity's Last Exam
#13 of 72
53.2%
DeepSWE
#21 of 22
28.0%
benchmark.space canonical public ledger of AI capability