Jan 142025 How should we test AI for human-level intelligence? OpenAI’s o3 electrifies quest Experimental model’s record-breaking performance on science and maths tests wows researchers.
