I am fed up of all this people that say DXOmark results are fake, normally these tend to be Canon trolls.
Do yourself a favour; go to dpreview and download the raws from D600, D800, 5D mkIII and 6D (if available)
Then you can process those files yourself as I did and you will se the huge difference in DR, specially at base ISO
you can see 100% crops of 6 cameras at iso 100 scaled to 25mpx (so they all have same size)
first row shows the original 100% crops and the second row shows same area with +4EV exposure
In my opinion the ranking would be:
1º C D800 2º A Pentax K5 3º F Sony NEX-5N 4º D Canon 5D mkIII 5º E 5D mkII 6º B Canon 60D
This test only shows DR at base ISO, and it proves DXOmark is right
By the way, I own a EOS 5DmkII and before that an EOS 5D, EOS 40D and EOS 1000D.
My second body now is an APS-C Sony NEX 5N as Canon APS-C sensors are a big piece of noisy crap