The time both OSes took to launch from startup is independent from xBench, right? I would guess that if xBench results is not to be trusted, time is time no matter what. Unless I'm missing something here...
This might be of interest to anyone following this thread. I posted in the Apple Rumors and Reports forum a link to an Ars Technica review of Leopard and one of the sections takes a look at OS X 10.5's performance.