DxOMark results on D600 out
Looks like the D600 is also based on the same generation sensor technology as the D800 and D7000.
(DxOMark D600)
If you look at something like the per-pixel SNR (what DxO calls "screen" as opposed to "print" measurements), then you can see clearly that the D800 and D7000 perform almost identically per pixel (since their photosite sizes are the same), and the D600 performing slightly better per pixel (since it has 6 micron photosites, as opposed to roughly 4 micron photosites on D800/D7000).
Of course, once you look at the "print" measurements, then the full-frame advantage shows, with the D600 performing almost identically to the D800, as expected. The D600 offers some interesting benefits over the D7000, for example. Consider that the D600 pixels are roughly 26% larger than the D7000 pixels, so linear resolution (in absolute terms) is lower than the D7000, BUT you have for remember that for a fixed print size (say 8x12), the D600 image will require 1.5x less magnification, i.e., it has roughly 19% higher final resolution in print.
All in all, I think the pressure is now on Canon like never before. They really need a significantly better sensor in the 6D (compared to even the 5DIII) if they want to remain competitive, since the D600 is almost one full stop better than the 5DIII / 5DII on many measurements. To illustrate: DxO's "low light" score is the highest ISO that yields a given SNR value (i.e., equivalent amount of noise in the image). On this measure, the D600 can be used at ISO ~3000, which yields the same amount of noise as the 5DII at ISO ~1800. For the 5DIII, the difference is more like 2/3rds of a stop, which is still significant.
I would love to see the 6D's sensor measurements, but I guess we will have to wait 'till the end of the year for that.