benchmark
.
space
benchmarks
rankings
compare
voices
transcripts
papers
articles
head to head
Claude Opus 4.8
vs
Inkling
10 shared benchmarks
9
wins
0
ties
1
wins
95.7%
AIME
97.1%
84.3%
BrowseComp
77.1%
80.5%
CharXiv
78.1%
91.0%
GPQA Diamond
87.2%
49.8%
Humanity's Last Exam
29.7%
77.8%
MCP-Atlas
74.1%
78.9%
MMMU Pro
73.5%
69.2%
SWE-bench Pro
54.3%
84.6%
Terminal-Bench 2.1
63.8%
76.2%
Toolathlon-Verified
45.5%