In the late 1990's, Precision Shooting published some tests done by Kevin Thomas when he still worked at Sierra. He incorporated some cryo-treated barrels into their routine bullet testing in machine rest guns. He found .308 stainless barrels lasted about 3500 rounds without cryo treating and lasted about 4000 rounds with it. The process increases wear resistance in stainless tools about 20% and that's more or less in line with what Thomas found. The effect on chrome-moly steel is greater, but Thomas didn't test cryo-treated chrome-moly; just untreated chrome-moly, and those lasted closer to 3000 rounds, IIRC.
Keep in mind that this is a tight accuracy standard. They could shoot about 1/4 moa with their special held-back lot of 168 grain SMK's and watch fliers begin to occur (the first sign of a shot out bore) very precisely. Shooters often can't tell the first few from wind or their own errors. Instead, the "fliers" might open groups to a little over 1 moa and stay that good for a long while.
I have heard Marine Scout-snipers report much longer life on their M24 5R barrels. However, as
this NRA article points out, they are much more interested in having a platform that keeps its zero through thick and thin than in having a benchrest knot-driller. They can be over 1 moa and still meet specs.
So this is a case of accuracy being subject to expectations.
This Shooting Sports USA article gives the .308 a bit more life (4500 rounds) than Thomas did, but I suspect that is partly a matter of shooters not being able to tell as precisely when best accuracy is really gone.
G. David Tubb suggests he can double the life of his barrels with his throat refinishing system. His accuracy standards are going to be more like those published by SSUSA. I don't know how that would play with a nitrided barrel. I don't expect it would do much with chrome lining.
The other thing cryo-treating is supposed to do is provide some stress-relief to barrels that weren't stress-relived originally. I've read that it does provide a bit, but its small relative to what you'd get from proper oven stress-relief. If you want a barrel that doesn't walk as it heats up, I would be looking at making sure the blank was properly stress-relieved before the barrel was contoured. This is most important for button rifled and hammer forged barrels, where process residual stresses are high.
Regarding barrel twist, this is a bullet length issue. Look at the bullet maker's recommended twist for the bullets you are using and stick pretty close to that. Extra spin causes two things. One you already mentioned, which is bullet disintegration. This usually begins with core stripping, which is when a bullet is spun up so quickly that the lead core starts to slip inside the jacket while it's still in the barrel. Even if such a bullet doesn't disintegrate, the uneven mass distribution that results will open groups up badly.
The other is the effect of bullet runout in the finished cartridges becomes exaggerated. If the bullet tilts even slightly in the bore, since modern pointed shapes have their center of mass (CM) ahead of the center of the bearing surface, tilt takes the CM off the bore axis. This results in drift off the ideal trajectory line. Where r is the length of the offset of the CM, the rate of drift is 2 pi times the spin in revolutions per second times the value of r. So if you have a CM a quarter of a thousandth off the bore line and the bullet is going 2500 fps, you have:
drift rate = 2 × pi × 0.00025 in × 2500 ft/s × 12 in/ft / 9 in
(Note that I am using 12 in/ft instead of the familiar 720 formula because I want revolutions per second, not per minute.)
That works out to 5.24 inches per second. So if the bullet flies 100 yards in 0.12 seconds, it will drift off the point of aim by
5.24 in/s × 0.12 s = 0.625 inches
So, if you randomize the bullet offset orientation in the chamber, then you get a group that is 1.25 inches wider than it need be. If you had 12" rifling, the addition would come down to 0.94".
So, pick your bullet and get the rifling to suit and not extra spin if you can avoid it. Some military barrels, like the 7" twist in an M16/M4 barrel is for stabilizing tracer bullets, which are long for their weight. But their accuracy requirements aren't stellar.