what should be the fitness function for these mutants?

I am working on mutation clustering where program generates 100 of mutants but for the cost reduction we have to select only the best and few ones,i have already tried it with k-mean, now i am applying GA algorithm here for example i have 1010 0001 1101 1000 0001 1001 0011 mutants, now what would be the fitness function for these mutants which provide me few mutants to process next step from this selection step.