Assuming Kiprop gets the win, it might be safe to bet on sub-3:50 at least.
I remember an article from Track and Field News not long ago about sub-3:50 miles, and how it would be very rare not to see one this year.
I couldn't find it again, but went to the IAAF world lists archives and since the first sub-3:50 mile was ran in 1975, the only years with out a sub-3:50 have been 1976, 1977, 1978, and 2006.
In 2006 it was close ... leader that year was 3:50.32 (Alex Kipchirchir Rono ... a guy with a 3:30.46 1500m PB, so clearly capable of better).
I did find a post at Jesse Squire's blog about all sub-3:50s and the metric equivalent (sub-3:33.00 apparently). Plenty of guys in Saturday's field have ran under 3:33 (some way under ...), but it seems the mile races are often more tactical.
Most go out somewhat quick, slow 3rd lap, then a big kick ... 3:48/49 for the win.
As to Kiprop ... I'm not sure he wouldn't want to make it a fast race. He just ran 3:26, so clearly if he goes all out or even close to that, no one will touch him in the mile either.
But, in a slower race, would he really be so sure of himself against Souleiman, just coming off a 1:42.97?