The only major, noticeable difference between RAM speeds is with significantly bandwidth or latency limited workloads - iGPU gaming, some database workloads, and a few other edge cases.
I spent an hour or two getting my Ryzen 5 Pro 4650G to run my (Micron rev. E) 3200c16 kit at 3800c16, and there's a very noticeable uptick in gaming performance there. But the overall system performance is the same. I also had a pretty easy time, just plugged in some numbers from dram calc and spent a tiny bit of time tuning some voltages for stability, and it's been 100% stable since. I was very lucky. Most RAM OC experiences are the complete opposite, with a lot of fine tuning and tweaking and overall a major time expenditure for small gains in the end.
AMD and Intel CPUs both gain overall performance going from, say, 3200c16 to 3600c16, or especially from something stock like 2133 or 2400 and to those speeds. But is it really noticeable in the real world unless you're actively looking for it? That's debatable. On my main system (5800X) I've just enabled DOCP (3200C14) and left it there. I have a vague idea of pushing it higher at some point as my RAM should handle it (Samsung B-die), but initial tests showed that even 3600 with dram calc numbers wasn't stable, and I wanted to actually use my system (!) rather than spend hours and hours tweaking RAM settings and testing stability.
My two cents: unless you have a workload you know benefits from the fastest ram speed possible, get an affordable 3600C16 kit or 3200c14 kit, enable docp/xmp, and leave it there. Splurging or spending tons of time tuning just isn't worth it - unless that's something you are specifically interested in, of course. You do you. But don't stretch your time or money for this imagining it's going to make a big difference.