The DR of the new E-M5 is easily the best of all Micro 4/3 cameras. That's a big deal. But even in this, the guess that the sensor was the same as the G3/GX1 was a good guess. The increased DR came primarily from the shadows, which can be achieved by lowering the noise floor, meaning that more meaningful data can be gleaned from the shadows amongst the noise. This is entirely in line with the assumption that the main difference between the E-M5 and the G3 wasn't the sensor, but the pipelines and processing. Just as the Panasonic GF1 and Olympus E-P1 showed wildly divergent noise characteristics.
Two other bits of data cast doubt upon Olympus claim. First, Olympus refused to say who was manufacturing the sensor. If Sony was actually the manufacturer, there would have been no reason to keep quiet. People would have been ecstatic. It is exactly what everyone wanted. The fact that Oly was remaining intentionally silent all but confirmed that the sensor was the same old sensor made by Panasonic.
Second, the noise performance of the sensor; if it was actually a Sony, it should have easily outperformed the new Canon G1X... but it didn't. Not at all. In fact, its noise performance was very similar to the Panny G3, as was stated in DPReview's workup of the camera.
The Panasonic's result looks worse but this looks like it is as much to do with the profiling (and lower black-point) than the Olympus, we suspect if we raised the black point to reduce the magenta tinge, the results would be comparable to the E-M5.These two points, and the fact that the one point for Olympus is tenuous, leads me to believe that the Olympus exec is wrong. I don't think that he's lying, since that would be too risky, but he may simply be misinformed. Lord knows, that happens a lot in large companies.
I hope that I'm wrong, but then the question arises, if the sensor is a Sony, why do the images compare poorly to other cameras? Olympus is usually a magician with their image processing. Perhaps they're just not used to working with the Sony sensor. If that's the case, cool. But I doubt it.