The Benchmark will be by far and away the least class A-sounding, so delete immediately from your list.
That's an interesting comment, since Class-A "sound" is far more influenced by design goals. A low feedback Class-A will have much different distortion products versus Class-A with more feedback, and that is not even taking into account voicing.
If someone makes a transparent sounding Class-A, i.e. neutral, not voiced, and no intentional distortion, then it would be closer to the Benchmark who has very low distortion at low power, somewhat negating any advantage of say Firstwatt and the justification for most ClassA, i.e. removing zero crossing distortion.