I've used both, in various formats over the past decade+; these are all posted about in my Crazybike2 and/or Delta Tripper and/or SB Cruiser threads, some have separate "review" threads too, if you need more details.
I thought I would really like an EM3EV A123 (LiFePO4 (LFP) pouch style pack I got (used), but it didn't have anything like the oomph I expected, even compared to a similar-voltage EIG NMC (LiIon (LI)) pouch style pack I also got used, on the same controller and motor. I expected the A123 to perform better, but at best it was the same. I don't really know why it didn't do as well; I never opened up the heatshrink to test cells and whatnot (other than to get to the BMS balance connnector to verify balance). It's been years since I used it so I don't recall the specifics of performance issues.
I have also had very large format Thundersky prismatic LFP cells, and they pretty much sucked in every way I can think of, but I expected that based on previous stuff I'd read that called them Thundersag. :lol:
I also got a well-used and already-broken 18650 LFP (unknown cell brand) pack, and it pretty much sucked too, but I also expected that since it was used and broken, even after I fixed what I could of it.
Lots of voltage sag under even "normal" loads, easy imbalance, etc.
I've had a small 18650 LI pack and it had issues, too, partly from a poorly-designed BMS that "locks up" and has to be reset, but mostly from cells that were simply not really capable of what they should have been from the specs the pack itself had (it heated up rapidly, too high internal resistance of the cells, possibly the interconnects as well but never tested those). I think that if the pack had been large enough with enough parallel cells in each group (only had 4p), it might've been able to do what it said it could.
I've used a small pouch-style LI pack in the frame of a Cemoto bike; it was ok for it's purpose, but it would not have been able to do the things I need to do in any of my everyday applications very well.
So far the only cells I have used and actually liked the performance and reliability of have been the EIG C020 NMC LI, which is unfortunate since they haven't been made in a decade or so.
Someday I'll have to try something new, and probably will end up with something from Batteryhookup or the like, some module out of a big EV most likely, adapted to my usage.
I expect it's likely to be a LI pack because most of the EV packs they get are that...but if they have a good well-performing LFP pack I can adapt, I'd be just as likely to get that.
Out of the many threads I've read and helped on for battery choices, problems, etc., I'd say that the main thing isn't the chemistry itself, it is the "cheap factor" (which is often reflected in the price, but even expensive packs can have the same problems--just less likely).
Going for the cheap(est) thing has tended to cause more people issues than anything else, as typically either the packs aren't made of well-matched equal-capability cells (sometimes actually recycled garbage cells, but mostly just no-name unmatched cells, sometimes claiming to be some specific brand / model, mostly not), or they're built poorly (causing high resistance in interconnections, or ICs that actually come apart under vibration/mechanical stress which disconnects some of the cells in some groups reducing pack capability and capacity).
Sometimes the packs are ok, but the BMS is a poor design or poorly made, and causes problems with the pack from one problem or another.
Sometimes it's both, so the imbalances created by mismatched cells is worsened by a BMS with no balancing function, or an actively-failed balancer system.
