You have Samsung "RAPID" enabled, don't you? Disable it and compare again.
That RAPID software is suspicious. It's not clear at all if it's actually any better than the Windows cache by itself. Windows will fill all of your RAM with data from the drive. It will pretty much never delete anything cached in RAM until some program needs more memory, only then something gets overwritten. If you think about that behavior, what's there to improve? It sounds like it can't work any better.
If you look up benchmarks of RAPID, you will notice that tests with actual programs do not run any better with RAPID compared to without. The only thing that gets improved is things like AS SSD.
I think what's happening with RAPID is simply that a program like AS SSD knows how to disable the Windows cache and test the actual drive, but it does not know how to do that for RAPID. I'm basically saying that RAPID is causing those numbers you see to be fake, AS SSD is not testing the actual drive at all.