This is an unscientific experiment, as it involves subjective evaluation.
I downloaded DPR comparometer pictures from 5 cameras at ISO 200 and ISO 3200 processed from RAW, loaded them in LR4, and used patch #8 for WB and equalizing exposure. Those were very slight adjustments. The cameras were: Sony Nex7, Nex5N, Nikon D3200, Olympus OMD-EM5, Panasonic G3. Then I cropped the pictures to 8x10' so they all looked the same size made glossy prints and gave it to my family members to sort them by pairwise comparison. Due to this cropping APS cameras had some disadvantage. During this sorting process, nobody knew which picture came from which camera.
Interesting that both ISO 200 and ISO 3200 were finally sorted almost in the same order, except the first and second place were swapped. This is an ISO 3200 ordered list, from best to worst:
1. Sony NEX5N
2. Olympus OMD-EM5
3. Panasonic G3
4. Nikon D3200
5. Sony NEX7
But I must add, that the differences are very subtle, all pictures look very much similar. Interesting that the "best" camera, namely NEX7 turned out to be the last in both cases. I think the main reason is visible moire in some places. It looks like NEX7 sensor has either too weak AA filter, or no AA filter at all.
Another interesting observation is that for whatever reason Olympus OMD-EM5 according to EXIF had much longer exposures than other cameras. In both ISO 200 and ISO 3200 sets, Olympus' exposure was 2.5 times longer than Nikon's, 2 times longer than NEX7, and 1.6 times longer than G3. I don't know whether it's attributable to different scene lighting or the real ISO on Olympus is significantly lower than on other cameras.
This has been extensively discussed on the forum already. The main reason for the difference in exposure times is simply varying light in the studio. DPR makes no attempt to keep the light constant at a perfectly controlled level. They have other ways to make sure that the results remain comparable.
Another thing you can notice the noise patterns and histogram profiles. They are very different for different sensors, but both Nikon 3200 and NEX7 as well as Olympus and Panasonic have very similar profiles, which adds credibility to the speculations that Nikon is using 24MP Sony's sensor and Olympus is using the same 16MP Panasonic sensor found in G3 and GX1.