Claude Sonnet 4.6 vs DeepSeek V4 Flash
6 shared benchmarks
4
wins
0
ties
2
wins
94.0%
AIME
94.8%
74.0%
BrowseComp
73.2%
89.9%
GPQA Diamond
88.1%
79.2%
MMLU Pro
86.2%
79.6%
SWE-bench Verified
79.0%
59.1%
Terminal-Bench 2.0
56.9%