From my observation, the longest acting benzos will cause less tolerance, because your'e not maintaining steady state by dosing something like alprazolam or etizolam 3-4 times per day. Dosing the short acting, fast onset variety will make it harder to perceive the effects of long-duration, slow onset benzos; this is because short acting, fast onset benzos will desensitize you psychologically to the gradual relief sought after from these substances. This is only taking into account usage patterns, as from my understanding the amount of times you re-dose will affect tolerance more than the dosage itself, within reasonable bounds(as in, medically prescribed limits).
The more complex issues unfortunately haven't been studied in detail as to which benzos have more selectivity for the different alpha sub-units. I can't recall which, but one specific sub-unit (a5?) is responsible for the physical dependence, whereas another (a3?) is responsible for the anxiolytic effect. Then you must factor in that some benzos have active metabolites which can cause a slower tolerance buildup, but which can accumulate if taken more than once or twice a week at moderate doses, resulting in accidental physical dependence.
Not all long-acting benzos are made equal when talking about accumulation either. You have benzos like medazepam with 3 or 4 active metabolites, diazepam with weaker metabolites than the parent compound, and exotic ones like phenazepam. Apparently phenazepam has a 10-16 hour halflife, but then metabolizes into 3-ho-phenazepam which has a similar potency(very rare for benzos to have equipotent metabolites), which extends its psychoactive effect to 60-90 hours; the only difference if you read the Russian literature, is that it's 3-hydroxylated metabolite is only a partial agonist at the BZD site and lacks any muscle relaxing effects.
It would be nice if the manufacturers actually gathered data on every benzo they release, and the sub-unit affinities, rather than just the overall potency. This might make benzos somewhat viable as anxiolytics if this information was publicized. The partial agonists, of which there are a few I believe, only show selectivity for the anxiolytic site, with maybe some negligible affinity for a couple of other sub-units. If one of these benzos lacked any agonism for the alpha unit responsible for muscle relaxation, sedation, etc, they may be viable to take for longer periods of time. As it stands now though, benzos are seen by the majority of the medical community as about as sophisticated as alcohol at this point, and will probably be slowly exiled from the medical arsenal, save for pre-op and very short term prescriptions.
Now, since benzos are basically playing with fire if you manage to take them a few too many days in a row, what's left? Well, there's gabapentin and pregabalin which work on calcium channels, and things like phenibut and baclofen that work on gaba-b. Gabapentinoid's and gaba-b agonists like phenibut build tolerance quicker for some people than even benzos! It seems anything affecting gaba can quickly downregulate its target receptor, but perhaps (don't take this advice seriously!) If you stuck to a precise regimen and rotated between a benzo, gabapentinoid, and gaba-b agonist while still sticking to a non-daily routine, you might be able to maintain near total efficacy if you need anti-anxiety medication as a life long crutch.
I won't get into the other options as I don't know much about them, but things like etifoxine and picamilon increase gaba output without any apparent risk of physical dependence. Picamilon is really interesting, as it's a gaba analogue that crosses the BBB, and then is oxidized to form plain old gaba and niacin. Both aren't available in North America, but are prescribed in several European countries for anxiety. It irks me that this is the case, as I actually do need to use anxiolytics every so often too, and the side-effects from benzos, pregabalin leave me feeling rather lacking in wit.
The whole problem with gaba in supplement form is it can't readily cross the BBB, and from my understanding Picamilon solves this issue. Whether or not it would be any better to flood your brain with more gaba, rather than using a modulator of its release, I really don't know. If you believe in the chemical imbalance hokey pokey, those of us diagnosed with an anxiety disorder have too little gaba available in the brain, so getting more gaba into the brain should solve the issue.
...I'll leave this post as is, I didn't mean to go on a tangent. For sure though, in other continents, there's less addictive options that supposedly work well enough for those who actually want anxiety relief, and not a recreational high. A fun fact if you read into the history of benzos is Halcion(triazolam), which one of the founding fathers of psychiatry was given during his psychiatric stay, was apparently known to cause dependence so quick it was quickly banned in several European Countries, yet in America it was still being prescribed outside post or pre-op situations. I'm pretty sure it's still used rarely to this day in North American countries.
