o1

Previous full o-series reasoning model. Trained with reinforcement learning for complex reasoning; the model “thinks before it answers”, producing an internal chain of thought.