Correct. The publisher in this case may have been misrepresenting a paid-access journal as open access, or the journal may have offered open access only to certain articles and presented this policy in a confusing way. I have certainly seen the latter before.