I really wish whoever performed those benchmarks would have gone the extra mile and kept the same machines/OSes in some kind of order.
You have one test with the eMac running 10.1.5, then running Jag, then the iMac running 10.1.5, then Jag.
A few tests down they reverse the machines, and also the OSes. It's difficult to see that the test proves anything, and really it doesn't prove anything. Jaguar is not all around faster, which surprised (and dismayed) me.
I don't have Jaguar yet, and with the many problems people seem to be having with peripheral devices working and networking issues (mail specifically), I may just wait until the update is out.