I've run some numbers in the past, though I don't have them in front of me anymore. While of course they should be taken with a rather large grain of salt, providing $6-9k/yr to every American doesn't seem to produce crazy numbers. It's also an amount that's low enough that we're not likely to see crazy numbers of people opting out of the workforce, while high enough to increase the bargaining power of workers a bit. If I had to pick an optimal level of UBI to institute nationwide (which in some senses must exist), I would guess it to be around there.

That said, recently I've been thinking of a different approach. Concerns about scale (and migration) are smaller if we start with a smaller amount. I've been thinking an approach might be creation of a private institution tasked with designing and coordinating a basic income program for regions (probably at the city or county level, in the US) that opt in and who meet whatever necessary standards. The program would then apply to anyone who lives or works in a participating locality. The program could then be gradually rolled out and scaled up as (/if) it's found to work. National benefits can be phased out by treating BI income as income. One thing that makes this plan seem more workable is that the organization could raise money from supportive individuals to build a buffer for the program, to serve as something of a carrot for prospective localities.