[eikenberry]

My guess is because it is a model trained specifically for security/hacking. So comparing it to Opus, trained for chat/code/etc., is apples-to-oranges.

[rs_rs_rs_rs_rs]

It is not, that's what surprised Anthropic employees too.