Performance Evaluation

Harvard Study Finds OpenAI’s o1 Model Outdiagnoses ER Physicians on Text-Only Cases

Researchers urge supervised trials before any clinical use.