The sensors are 4k x 3k with 2.74 micron pixels, and 6.5k x 4.8k with 3.45 micron pixels. They are backside illuminated, which apparently mitigates some of the problems of smaller pixels and increases sensitivity.

I’ve recently been using two new Sony Pregius sensors (needed for the global shutter). The 12K BMD camera has 2.2 micron pixels. The Alexa has 8.25 micron pixels, which appears to be a sweet spot, at least for an earlier design that has aged quite well. The old Kodak CCDs, which looked great, had 5.5 micron pixels. I’d guess that extreme cropping would just reveal the (carefully chosen) faults of the lens - so that wouldn’t be an advantage (and I suspect most people here will be nervous about the potential of a producer deciding to completely reframe your images). But what is the effect of a higher resolution sensor (even with pixels that are too small) when used with a vintage lens or diffusion-not looking for resolution nor shooting test charts? You’d get the theoretically better color science of this camera/sensor and less aliasing. With photosites that small we are really at visible light diffraction limits. Even more crazy is anything above f8 on that lens would be diffraction limited lower than the resolution of the sensor. Tommy wrote: To test the resolving power of the filter array would require a lens rated at 227 LP/mm or better with a ~30mm image circle.

Zeiss siemens star test chart