The exit polls are looking fairly bad for the (R) side right now, however, unless the exit pollers have considerably improved their methodology the exit polls will still have about a 5 percent systematic error to the (D) side.
If so, here's what the numbers look like (using numbers from The Corner):
Virginia (52-47) prediction very close*
Rhode Island (53-46) prediction (R) (!!)
Pennsylvania (57-42) prediction (D)
Ohio (57-43) prediction (D)
New Jersey (52-45) prediction very close (!!)
Montana (53-46) prediction (R)
Missouri (50-48) prediction (R)
Maryland (53-46) prediction very close ---
but I think it will be a clear (R); remember this is all based on someone's model of the overall election. If the African-Americans go just a little bit more (R), this might turn.
Tennessee (51-48) prediction (R)
Arizona (50-46) prediction (R)
Update: I cleaned up the format a little bit, mainly using a monospaced font to make things line up without having to actually create a table. But I note one thing: I made an arithmetic mistake on Virginia, which I showed as "very close" where it should be a clear (R). Since then I've seen considerably more results, so I hate to change the original, but by the method, that should actually predict Allen by at least a couple poiints.