Minimum spanning tree with small set of possible edge weights

Given a undirected graph which only has two different edge weights $ x$ and $ y$ is it possible to make a faster algorithm than Prim’s algorithm or Kruskal’s algorithm?

I saw on Kruskal’s algorithm that it already assumes that the edges can be sorted in linear time with count sort, so then can we gain any benefit by knowing the weights of all of the edges ahead of time?

I don’t think there is any benefit, but I am not sure…