benchmark
.
space
benchmarks
rankings
compare
voices
transcripts
papers
articles
head to head
Claude Opus 4.7
vs
DeepSeek V4 Pro
10 shared benchmarks
6
wins
0
ties
4
wins
79.3%
BrowseComp
83.4%
94.2%
GPQA Diamond
90.1%
96.3%
HumanEval
76.8%
83.0%
LiveCodeBench
93.5%
87.0%
MMLU Pro
87.5%
59.2%
MRCR v2
83.5%
85.7%
SWE-bench Multilingual
76.2%
64.3%
SWE-bench Pro
55.4%
87.6%
SWE-bench Verified
80.6%
77.0%
Terminal-Bench 2.0
67.9%