It depends what you mean by big-bang. I consider the big-bang to begin with inflation, not with a singularity, so that the starting point is the inflating universe, making no hypotheses about what came before (if the question even makes sense).
The inflating initial starting state is for all intents and purposes, a perfect deSitter state which is adiabatically growing as the inflaton slides down the potential. At the end of inflation, when the inflaton starts shaking non-thermally, the state is no longer unique, but the semiclassical description of the initial state is by a thermal state inside a deSitter horizon.
The natural entropy to associate with this state is the area of the cosmological horizon in Planck units, and this entropy is far from zero. But it is infinitesimally small compared to the maximum entropy we could squeeze into the universe today, given that the cosmological horizon has grown so much, but past the end of inflation, the growth has been out-of-equilibrium.
So the entropy of the initial state of the universe is about the square of the de-Sitter radius at the end of inflation. I don't know a precise number, but suppose it's a deSitter temperature of $10^{14}$ GeV, that's about a million planck lengths of radius, so a dimensionless entropy of order $10^{12}$. Compare with $10^{135}$, which is the maximum entropy you can squeeze in the modern cosmological horizon, and you can see how low-entropy the initial state was, despite being in thermal equilibrium at the time.
This explanation of the low-entropy initial conditions requires you to consider a single horizon-volume as all there is, and this is the holographic view of inflation promoted by Banks, Fischler, Shenker and Susskind. It was suggested to be the reason for the low-entropy initial condition by Davies in the early 1980s, but it is still not accepted by the astrophysical community, for reasons that I wouldn't be able to properly explain, because I think they are ridiculous.