Sensor size plays nearly no role in the final product. It’s crazy that many people still don’t understand that because there are countless examples. Bad image processing, bad exposure time, artifacts, bad autofocus, bad lens, there are so many factors. A larger sensor size does basically nothing else than making it easier to produce lenses with a larger effective diameter for a given field of view. A sensor can only capture the light that goes through a lens. A larger effective diameter captures more photons per time from an object, this reduces the relative deviation of the number of photons that arrived during the exposure. So the raw data is less noisy. A smaller lens achieves the same effect with a longer exposure time. And a small sensor captures exactly the same amount of light per time from an object as a large sensor, if you use exactly the same lens. There isn’t anything magical about a larger sensor.
Image quality on the Pixel 5 is lower than on the Pixel 3. Full quality example pictures from Android Authority demonstrated this. The Pixel 5 uses more severe noise reduction and has more highlights that are blown out. Additionally, the Pixel 5 captures less light from bright places than the Pixel 3 due to bracketing. Additionally, Night Sight can take up to 6.5 seconds, which frequently causes images to be slightly fuzzy. The Pixel 3 boasts HDR+ improved, a quick Night Sight mode that produces far crisper photographs. By the way, in HDR+ On mode, the Pixel 3 captures worse daytime images than the Google Nexus 5x.
Analog pixel binning is older. Dual pixel sensors also use pixel binning, that’s why Canon Dual Pixel raw files are twice as large. Dual pixel sensors with 12 megapixels do actually have 24 megapixels. Also, there are Bayer sensors without Dual Pixel autofocus that offer analog pixel binning or at least a combination of analog and digital pixel binning. Then you get 3 megapixels instead of 12.