Because the N in SU(N) doesn't refer to the number of basis states. It is linked to the number of spatial dimensions of the world (for how that link works, I recommend taking
@weirdoguy's advice and learning about why SU(2) is the double cover of SO(3) and why, when we include spin-1/2, that makes SU(2) the correct group for representing spatial rotations), and that doesn't change when you look at spin-1 particles instead of spin-1/2 particles.
What
does change when you look at spin-1 vs. spin-1/2 particles is the
representation of SU(2) that you use. Heuristically, for spin-1/2 particles you use the representation of SU(2) that uses 2x2 matrices, whereas for spin-1 particles you use the representation that uses 3x3 matrices. In other words, the size of the matrices in the representation is what refers to the number of basis states. (There is a lot more here that I am sweeping under the rug, even though you labeled this as an "A" level thread; a real "A" level discussion of this topic would take a book, and there are indeed plenty of them.)