My test setup was my modest desktop computer running an AMD Athlon XP clocked at 2.0 GHz, with 1 GB of RAM, running Ubuntu 8.04. Tests were run on Windows XP SP3 with the latest patches running on Sun's VirtualBox with 192 MB of RAM made available to the virtual machine.
Scores for the Acid3 Test - Best to Worst
Safari 4 (Windows Public Beta) - 100
Firefox 3.5b4 - 92
Opera 9.6 - 85
Google Chrome 18.104.22.168 - 79
Internet Explorer 8 - 20
Internet Explorer 7 - DNF
Scores for the V8 Benchmark - Best to Worst
Google Chrome 22.214.171.124 - 1444
Safari 4 (Windows Public Beta) - 1058
Firefox 3.5b4 - 147
Opera 9.6 - 98.9
Internet Explorer 8 - 29.8
Internet Explorer 7 - 16.6
Based on these results, I'm giving the win to Safari 4 with Google Chrome coming in second. Both of these blew the doors off the competition in the speed department. Chrome leaves something to be desired when it comes to standards compliance though. It's also no surprise that Google's browser excels in the benchmark written by Google.
I think standards compliance is more important than speed. Slow and correct is better than fast and incorrect. Based on that, one could argue that Firefox 3.5b4 deserves second place. I'm hoping that the Firefox developers can work out the last few kinks to fully pass the Acid3 Test for the official 3.5 release.
It's also noteworthy that IE7 and IE8 both made a dismal showing in these tests. Perhaps this poor performance is by design. Based on these results, Microsoft could certainly make a case that including Internet Explorer with Windows is not an anti-competetive practice. It's certainly non-competetive, but I don't think it's anti-competetive.
For reference I'm providing screenshots of the results of each test.