GPT-5.5 Instant Outperforms Doctors in Medical Knowledge

OpenAI's GPT-5.5 Instant is now outperforming doctors in medical knowledge evaluations. This new model achieved superior scores compared to physician-authored responses in areas like communication and precision. What's interesting is that factual errors in medical responses declined by 71% over an eight-week period. OpenAI says its ChatGPT platform now attracts 230 million weekly users seeking health and wellness guidance. Karan Singhal, a former Google Med-PaLM architect, is leading OpenAI’s medical AI initiatives. His team developed a specialized assessment framework called HealthBench, collaborating with over 200 practicing physicians. The company's ChatGPT Health feature, which includes medical record integration, debuted in January and currently has a waiting list for access. This matters because it shows how AI is rapidly advancing in sensitive and critical fields like healthcare.