Are color mix nodes efficient?

What happens inside this node? I know we can use it for float operations, but is it advisable compared to using math nodes (custom even)? Is it slower to do 1000 white+white (3ch), red+red (1ch), compared to 1+1 math node?

I mean, does it serve an efficiency purpose to do a custom Lerp group with math nodes compared to a single color mixer?

The color mix node can do various operations that are hard to do or is not possible to do with math nodes (basically using it to ‘cheat’ at math if you will, not like it will harm the final output anyway).

You also need to note that color mix nodes operate on all channels at once. Otherwise you would need 4 different sets of math nodes (one for each channel) bracketed by separate/combine RGB nodes.

It is true that math nodes are indeed more efficient and practical for some operations, but color mix nodes can also have some good use.

Yeah just noticed. Fed fresnel float into a normal input, passing it through a group untouched, and out came a gray fresnel, meaning it was converted to three channels. I just feel it’s wasting something when all I’m dealing with is a float. Example of a normal linear interpolation as a float group; makes up a node group and consists of 4 math nods, compared to a color mixer which does the same thing as a single node (but expected to calculate same thing on all three channels):

Which one is most beneficial wrt rendering speed? I btw hate using the color mixer with custom value inputs as I can’t limit their input same way as I can with values exposed to a group interface, and letting them stay as colors makes them unreadable until you click on the color making maths difficult to debug.