I think especially in math, you probably have to go up eventually. We've tried to go "deeper, not faster," but it's not easy to do with even one child. Trying to differentiate for a child this GT is very, very hard in a regular classroom. Even grouping is unlikely to help much, since fininding even one of these kids in a class is rare; finding two is statistically improbable.
I think at least differentiation would be necessary, and frankly, I don't think most teachers are that dedicated to differentiation that they'd be willing to do what it would take. There are some fabulous teachers who are good at differentiation and use it consistently, but they are not to be found in every class...or even more than a few classrooms...
And I guess I'd argue that good differentiation is going to result in a form of acceleration. Even if the child is physically in a same-age classroom, they're being taught as if they were in a higher-level classroom. If it's working, they're going to advance each year, getting further and further ahead of their agemates.
(Of course, this sort of consistency in differentiation over so many years is rare in the extreme! Usually a kid is lucky to get one good year of differentiation. Two in a row is practically a miracle, from what I've seen!)
So does that mean I vote a wimpy "no"? I think it might.
