What are you measuring, comparing 10.6 on the Mac Pro 5,1? Is it comparing 10.6 to newer OS releases or to other systems?
Once you have solid state storage, you can only launch TextEdit or list folders in a file so instantly, regardless of what else is in a machine.
This could be different on the Mac Pro specifically, but my experience on 2007 and 2011-era MacBook Pros was that in general, the newer OS releases themselves work faster with what was already in the machine, especially 10.11 which fixed a few long-standing memory leaks I'd been seeing on my machines.
This of course wasn't really measurable in benchmarks, at least the ones I was using. In my experience, measured application performance doing things like producing a render almost never differs from one OS release to another, unless you start talking about major hops, such as 9 to X or 98/Me to XP and 95 to NT4. (This is especially true because suitable benchmarking conditions usually involve a relatively clean OS image and a fresh warm reboot, so benchmarks usually avoid problems with memory leaking or caches filling.)
