The High ISO performance between two competing bodies of the same generation are very close, which means two things:
1) The differences aren't that important in actual, practical use.
2) To measure those small differences you have to use an exacting methodology. This means the same absolute exposure (aperture/shutter speed) with the same lighting. Matching the aperture requires using the same lens on both bodies, which for Canikon means using a Nikon-mount lens with a manual aperture ring, and adapting the lens to the Canon body using a F-mount -> EF mount adapter. Failing to use the same lens on both bodies means the actual aperture/t-stop can vary by up to or more than a 1/2 stop, which is larger than the actual High ISO difference between most bodies of the same sensor size. Naturally you also have to use a tripod, timer release, MLUP, and identical framing.