Claude Opus 4.8 vs Kimi K2.6
9 shared benchmarks
7
wins
0
ties
2
wins
95.7%
AIME
96.4%
84.3%
BrowseComp
83.2%
80.5%
CharXiv
80.4%
91.0%
GPQA Diamond
90.5%
49.8%
Humanity's Last Exam
34.7%
78.9%
MMMU Pro
79.4%
54.8%
OSWorld 2.0
22.1%
69.2%
SWE-bench Pro
58.6%
84.6%
Terminal-Bench 2.1
71.3%