Indeterminate. If the atoms form a perfectly mixed solution then
you might guess 2a+4b where a and b are the fractions of A and B.
But if they form a super-lattice where the stacking of the atoms
only repeats over a long range (as happens in the many structures
of silicon carbide [silicon and carbon are not metals]) then you
can get almost anything, with various seemingly unconnected
sequences and lattice lengths, even in different directions. For
very dilute things like A50B you will get the normal lattice period
of A over much of a crystal but with lattice distortions around the
occasional B atom. That would cause a broadening of x-ray
diffraction patterns.