A single day holds at an absolute maximum of 3^30 different combinations but in practice it's probably much less than that because people converge to more green boxes (generally) as they guess more. I'd bet by 3 days though you've provided a unique set of answers.
That's the ballpark I was thinking, too. The only wrench I can think of that might get thrown into it -- there are some known popular starters and popular guessing styles. If you go for, say, ARISE and then hunt vowels, I bet the number of collisions could be strung along a bit longer.
Anyway, that's a pretty constructed scenario, but it is sorta interesting to think about.
In any case -- fortunately it is NYT, so I bet they'll happily just let it go to a nice stable daily crossword sized population and stick there indefinitely without messing it up. Maybe they can add a 6-wordle for subscribers.