A preview version of OpenAI’s o1 reasoning model achieved higher accuracy than hundreds of board-certified doctors in six diagnostic and management reasoning experiments, according to a *Science* ...