How big are the differences in performance and reliability - speed rates being equal - between a DRAM nvme gen 4 to a DRAMless (HMB) nvme gen 4 ? In other words, if you're the average user and could get a DRAM drive for $50, or an HMB version of that same drive for $35, which one would you go for?
With some noteable exceptions (like Samsung's recent problems with 990 and 980 PRO NVME's) I believe SSD's have been very reliable as a whole. But don't confuse reliability and service life since I believe the more important thing to look at is life, or how long they last in use. You can use TBW (TeraBytes Written) spec's by manufacturer as a guide for that since their warranties last up to the TBW spec, or the number of years, whichever comes first.
Drives with QLC flash have relatively low TBW specs and since QLC flash is cheaper it makes sense you'll find it in DRAM-less drives. That doesn't mean it's bad though, it just means you have to consider your useage pattern. A game store drive, for instance, gets very low write useage as it's mostly reading data making a QLC drive a good fit. If you do extensive work that involves constant or frequent writing of very large files, video rendering as a content creator perhaps, you might want to choose a drive with a high TBW spec to provide long life and would probably use TLC flash. That doesn't mean a TLC drive will automatically have DRAM cache though.
I would not suggest a QLC flash drive for the system drive as it gets so much write activity. Also, if you get a DRAM-less drive that uses an HMB cache instead be sure to also go to the manufacturer's web site and download their NVME driver to be certain the HMB is enabled.
I recently upgraded my system's drives and went with TLC drive with DRAM-cache (980 PRO) for the system drive, 512 GB to keep cost down, and a Crucial P3 Plus DRAM-less QLC drive for the games store. Because it was QLC and DRAM-less I could go with 2TB and it's turned out to be a perfect trade-off since I can't really tell any difference in performance.
And reading through reviews as I made my selection I came to realize there's are significant differences in performance even of drives with similar specs and using similar flash and controllers. I think what it comes down to is how well-tuned the algorithms are that the drive manufacturers use. But even then once you get out of the synthetic benchmarks, like Crystal DiskMark, used by the reviewers you'll not likely notice any difference while operating the computer except in highly specific scenarios.