On K-stability, height bounds and the Manin-Peyre conjecture