Hugh difference between benchmarks and actual power. It's really easy to make a mobile chip give a high benchmark score, it's really hard to sustain that level. Opposite for CPU's
this means nothing the differences between what an arm chip is doing and what a computer CPU is doing is so vast as to make these comparisons pointless and misleading.
This is like running a passenger car against a semi and going "Holy crap look how much faster the car is, Eddie Stobart are now going to swap their trucks out for Nissan Micras"
Cinebench uses closed-source tests which makes them completely pointless for cross-platform comparison. Most you can do is compare hardware running on the same platform.