One could say that down to a markov chain with noise, the perspective is the new model named after the problem solves it significantly more reliably than the previous without a problem specific hack.

It's also worth noting the current model is the lower scoring o1-preview, not o1.