Access to the o1 model expanded

Expanded access to o1 model. The o1 series of models are trained with reinforcement learning to perform complex reasoning.