head to head
Claude Haiku 4.5 vs DeepSeek-V3.2
1 shared benchmarks
0
wins
0
ties
1
wins
80.7%
AIME
93.1%
benchmark.space canonical public ledger of AI capability