32 or 33 every year, as best I can tell. The reason for the uncertainty is because of the subset of the entire test that counts towards a passing score. The scaled score is for the entire test. The people taking the exam for real have the "trial" questions taken out of the equation. If you did really well on the trial questions, it could boost your scaled score(to 32 for instance) but when those questions are removed for the scoring of the actual ABA written portion, you could potentially drop below a passing score.
This is all theoretical based on many years of observing scores. I have never seen a 32 scaled score NOT be a pass for the recent graduate taking it for real, but I think it could happen in the perfect storm scenario described above. If you get a 33, it is probably a sure thing.
I hope this answers your question.