During the old space race days, launching double or multiple scientific probe copies seems to have been the standard. It should have the advantage of sharing the one time fixed development costs while less than doubling the manufacturing and operational costs given diminishing marginal costs of reiteration.
Based on Pioneer, Viking, Voyager and other duplicate missions (even to some degree Galileo/Cassini and MSL/Mars2020 and the Soviet Venus missions), how should one estimate the marginal cost today of duplicating a Cassini class orbiter to Uranus with a copy to Neptune at the same time (albeit on separate launchers a year apart)? Would major design differences of the spacecrafts and their instruments be worthwhile to turn it into two unrelated missions?