In the standard model, the quark masses are proportional to their coupling to the Higgs field; but the magnitudes of those couplings (called yukawa couplings) are unexplained.
A theory which goes beyond the standard model, by explaining why the values of the yukawa couplings are what they are, has the potential to explain why the tilt of the quark masses is different in the first generation, compared to the other two generations.
For example, in some "vacua" of string theory, the yukawas depend on the surface area that quark strings have to cross in order to interact with the Higgs string - as the area to be crossed increases, the yukawa will be exponentially smaller. The distances and areas in turn depend on the resting configuration of the extra dimensions, and of any branes that they may contain.
A model like that has the potential to deliver an explanation, e.g. in terms of how the various branes arrange themselves when at rest. Although in practice, these geometric equilibria cannot presently be calculated in anything like the necessary detail, and it is counted a success if one can just show that one quark (the top quark) will be much heavier than the rest.
For now all this is speculation. Even if one believes that string theory is definitely the theory of everything, one has to admit that it encompasses vastly many possibilities, and we do not know at all which of them corresponds to our world.
I have, over the years, seen one or two theories of mass in which the tilt of the first generation had a clear and specific explanation. I will link them here if I can find them again.