I like to think of the turing test as merely a small fraction of a greater benchmark, ideally you'd have a big table with say, 20 different tests you could try your system on.