can you add in the other quants like IQ3_M?

also my personal simple rule of thumb for local ai sizing is:

max model size (GB) = ram (GB) / 1.65