Supra's go to big singles because they are more simple to plumb as well as the fact that there just isn't room for bigger twins. The fastest 2011+ turbo Mustangs are running a mix of single and turbo twin turbo setups. They are making 1100-1300rwhp so both can work.
Smaller twin setups have some advantages. They can be placed closer to the exhaust manifolds, which helps to enhance spool (the further downstream the turbos, the more heat is lost, and cooler exhaust loses volume and energy to spool the turbos with). Twins typically get equallized exhaust flow since they are typically placed the same distance away from the exhaust manifolds, where as a single typically has a short exhaust routing from the bank closest to it and a longer run from the other side of the engine.
Disadvantages are that there is more to go wrong with a twin setup. Two turbos, two wastegates, two oil lines, two coolant lines, etc. Twin setups are also typically heavier (more overall piping and the weight of a second turbo). Twins are also typically more expensive as you are buying a second turbo.
With the Platform, and specifically the 3.5 V6, part of the magic I believe is the packaging of the turbos so close to the exhaust manifolds. In fact, Ford was so intent to do this on the 2.0, that they integrated the exhaust manifold into the head casting. You can't attach a turbo any closer than that. The proximity to the exhaust in conjunction with size of the turbos work to help create the instance spool and torquey nature. Moving to a larger single setup should be reserved for the situations where a larger twin setup won't fit for packaging reasons unless you are looking at a race only application.
As far as a larger turbo making more power at a given boost level vs a smaller turbo, things have pretty much been covered, except for understanding compressor maps. It has somewhat been covered in noting that a larger turbo typically has lower intake temps vs a smaller turbo at the same measured boost level. When comparing turbochargers, and any forced induction application, one should look at the compressor maps and try to match up the compressor map to the intended operating range of the turbo or supercharger. If you go too big or two small, you will operate outside of the peak efficiency range and will see less than ideal results.
With the upgraded ATP turbos, I don't think you are going to see any updated compressor maps, since they are upgrades to an existing turbo. As manufacturers start coming out with new turbos, we'll want to pay close attention to compressor maps on top of all the other variables.