A few issues here:
1. You are assuming that the number density can be calculated by knowing the volume of an individual item  in this case a molecule  like this:
N = Unit volume/volume per object
But you are using an expression for the volume of an individual molecule that assumes it is spherical, like a tennis ball, and spheres can not be packed perfectly together. Thus the number of tennis balls that can fit into a volume v is not n = v/(pi*d^3/6), but is actually less than that due to the interstitial holes between adjacent balls.
2. Further, the values of n1 and n2 are the number of molecules per unit volume. That number is certainly much greater than 1. So no way n1 + n2 = 1. As a check, suppose d1 = d2: in that case your formula gives you n1+n2 = 1/(pi d^3/8), which is not 1 unless the volume of a single molecule is equal to a the unit volume.
What may be of use here is that if you know the values for n1 and n2, then if you mix these two types of objects together in a unit size container (v1+v2=1), then the total number density is:
N = v1*n1+(1v1)*n2
This is basically the weighted average of n1 and n2.
