It isn't in the metric that really matters for day-to-day usage which is disk IO.

The 5 & 4A have the same GPU and the 5 has a slightly faster CPU. Unclear why the benchmark scores are what they are, possibly a worse thermal profile with the 5, but if you're not gaming the faster storage will have far more impact.

Pixel 5 owner here, I am disapointed at that metric, but I am not disappointed at the speed of my phone. I can see how that would cause others to second guess a Pixel 5 purchase or have buyer's remorse upon learning that.

In most benchmarks they are basically within the error margin, but I have indeed seen a few where Pixel 5 does strangely poorly. I am definitely curious to know what's causing that, Pixel 5 has a slightly newer version of the same chipset (765G/Adreno 620 vs 730G/Adreno 618). The two best guesses I have is either the waterproofing really messes up with the thermals, or there's some unlying bug causing bad performance in those specific benchmarks.

That being said, in real-life use, unless you're playing some heavy games, phone devices are generally more memory bound than CPU bound in my experience, so the 2GB extra ram will go a long way.

Waterproofing or not, phones are almost always effectively hermetically sealed from a convection point of view and do not rely on convection in any form for cooling. I do not see how waterproofing would interfere with conducting heat away.

Maybe it forces a different internal layout?

The pixel 4a 5g uses the same cpu as the 5, does not have waterproofing, and has the same benchmark speeds as the 5. I think the layouts are different too, based on the fact that the 4a 5g is larger than the 5.

I'm fortunate enough to have needed a Pixel 5 _in addition_ to a Pixel 4a, results are nonsensical imo, 5 is _definitely_ more buttery

I think this is the real point; 90hz is way more important than CPU speed in making it feel like an upgrade.