I've been measuring impact frame data for SC1 beginning with Hwang.
For standing A,B,K I'm getting the following:
A
AC: i9-10
DC: i9
K
AC: i9
DC: i7
B
AC: i10
DC i10
In testing, A clashes with K on both AC and DC versions, so presumably they are both i9? B is about 1 frame slower, so...