There's no real benefit to one technique vs the other that I know of. In either case, the usual potential problem with SendInput is that the input focus could change between when you send the input and when the input is received by the target application. The classic example is a modal dialog appearing asynchronously to alert the user of some problem, and ends up stealing focus, and soaking up the remaining input that was supposed to go elsewhere. There's no difference between doing a successive series of SendInputs vs a single SendInput in this scenario.
(I'm pretty sure that internally, SendInput is just looping over the inputs passed to it anyway...)