ChatGPT Health goes wide with 300M weekly queries — and a fresh lawsuit

OpenAI is flipping the switch on ChatGPT Health for all U.S. adults, pushing the feature out of a testing hub and into every chat — just one day after a Florida pastor sued the company, alleging it gave advice that nearly killed him.

The timing is awkward, to say the least. The company shared that users are now firing off 300 million health-related queries per week, up from 230 million in January. Starting this week, anyone on a free, Go, Plus, or Pro plan can link data from Apple Health, MyFitnessPal, and medical records systems like Epic and Oracle Health. The big change? You no longer need to visit a dedicated health hub. Seventy percent of health queries during testing already happened in the main chat interface, so OpenAI is simply meeting users where they are.

On the model side, the company claims its smallest new release, GPT 5.6-Luna, now beats the larger GPT 5.5 on HealthBench, its internal benchmark for evaluating LLMs on medical questions. OpenAI stresses it doesn’t train on user health data and works with physicians to tune responses. But those performance claims sit uneasily next to the fine print — the terms of service still explicitly state the tool isn’t meant for diagnosis or treatment. The company pointed to that disclaimer in its response to the lawsuit and told The New York Times it’s working on safer medical answers.

The tension here is real and unresolved. Studies keep showing AI bots aren’t reliable for medical advice, yet the industry keeps shipping. Anthropic and Google are all-in on health AI features. OpenAI’s rollout makes the bet that 300 million weekly queries represent a demand too big to ignore — even with a lawsuit that alleges the stakes are life and death.