In principle, a sensor that measures full color at each pixel location can deliver better image quality than a single-color-per-photosite sensor that relies on demosaicing. It can preserve more true detail for a given pixel count and reduce color aliasing artifacts, which can also lessen the need for an anti-aliasing filter and improve sharpness.

However, in practice these multi-color designs have tradeoffs. The main drawback noted is poorer light sensitivity, which hurts low-light performance. That makes them less practical for many mainstream cameras, where good high-ISO performance, efficiency, and overall versatility matter a lot.

That’s why most modern cameras use the more common single-color-per-photosite approach: even though it requires interpolation to reconstruct full color, it offers a strong balance of image quality, sensitivity, and practicality. The color channels in real images are also correlated, so interpolation is less damaging than it might seem at first.

So the short version is: multi-color-per-pixel sensors can be superior in theory for detail and color accuracy, but single-color-per-photosite sensors are more practical and generally perform better in low light, which is a major reason they dominate the market.