I think what you'd find is that there is no consistent percentage that corresponds neatly to the way we currently divide species. Some species are a fair bit more diverse than others, and species barriers (i.e. cessation of viable reproduction) can arise from just a few mutations.