As far as I'm aware 1mg xanax (alprazolam) 2-3x per week would not cause withdrawals... Like 2-3x a week is not even a "habit" for a drug with a duration of action measured in hours, so you wouldn't be "quitting" if you stopped doing this. I mean I guess you would be technically but this terminology is usually reserved for substances taken daily, or almost daily, at least, such that when you wake up you're already "withdrawing" slightly...
(EDIT: I'm not a doctor, the above paragraph isn't medical advice, anyone reading this thinking I'm saying using alprazolam 2-3x a week is 100% safe, I quite explicitly am not. I'm just saying that relatively speaking - this does not sound that bad - to me! obviously... YMMV! Same goes for anything else I've said here, the dangers depend on a lot of factors, I feel like you probably realize that already but since a few people have commented that even this relatively miniscule level of usage may not be entirely safe I feel the need to add that this is also true...)
"Dependency risk" is a bit harder to quantify, alprazolam is not a benzo that's ever really appealed to me but that said I'd say the dependency risk is higher. Actually that's a lie that it's never appealed to me, I find alprazolam extremely compulsive, it just hasn't been "habit forming" for me personally because when I have taken it it's generally been SO compulsive that it makes me do enough weird shit in a short enough time to put me off taking it ever again. Not because I enjoy taking it as such, mind... actually maybe that's a lie too, I just don't remember enjoying it but I must have been getting something out of it while I was on it for it to be compulsive, it just has way too much of a sobriety delusion for me that I escalate dose way too easily.
Phenibut I think has a much lower risk of dependence just because you basically cannot take it multiple days in a row without escalating the dose ridiculously, if you're a person that thinks about that kind of thing, which it sounds like you are.
Generally speaking faster onset and shorter duration means greater addictivity and a higher dependency risk, phenibut has quite a drawn out duration, and lasts quite a long time compared to xanax which hits quickly and fades quickly as well.
I can't (or shouldn't) speak biochemically because I don't have the credentials to do so, beyond that I don't know by what metric klonopin (clonazepam) can be said to have a greater dependency risk, a lot of this depends on the person but I don't personally consider clonazepam to be particularly dangerous, or habit forming. On the other hand I've taken a lot more clonazepam than I have taken alprazolam, for that exact reason, but my usage is usually very premeditated, I'll take between 2-4mg a day for a few months at a time then taper off. I know people talk about kindling and all that shit and I'm not saying this is gonna be true for everyone but I just don't seem to have much of an issue jumping off from about 2mg/day after months of staying pretty solidly within that range, the most I've escalated up to and then jumped off cold turkey was 6mg/day which I'd say was not advisable and I don't plan to do that again but it was a stressful time and just bad planning on my part suddenly running out.
To give some context prior to ever messing with benzos I was taking phenibut, 2-2.5g / day, 2-4x week, for months at a time on and off for, quite possibly almost 10 years. I'd consider clonazepam to be definitely sketchier than phenibut for aforementioned reasons - it's just much easier to escalate your dose, and typically when I do take clonazepam now I'll take it essentially daily for a while which I did not do with phenibut, also for obvious reasons.
Generally diazepam I believe is still the benzo-of-choice for tapering from long-duration dependence because it has an extremely long duration of action, as far as I'm aware, so on that basis clonazepam should be "safer" than alprazolam. I'm not sure how to look at that pharmacologically in terms of GABA-A downregulation and whatnot but I'm not sure it really matters, in this context... Unless you have some specific reason to be thinking about this stuff or just find it interesting (which is valid) IMHO when you've already decided to take SOME kind of benzo, or other class of intoxicant, and start obsessing about receptor binding affinities and the like, it's mostly just adding an unnecessary layer of obfuscation that most of the time just serves as a kind of post-hoc rationalization of whatever subconscious decision you've already made, so you can say to yourself well of course X drug is a better decision than Y drug because of this activity at that receptor... whereas in reality the neuropharmacology is far more complicated than that and (again, in this specific context, comparing benzos) not really going to outweigh your own individual neuronal wiring.
Of course longer duration benzos cause more neurological changes, be in GABA-A downregulation or something else - very broadly speaking - but that's because you spend more time under the influence of the drug. It's not quite an apples and oranges comparison but it's also not quite apples and apples.
