OpenAI expanded ChatGPT Health to all US users over 18 across all subscription tiers, the company said today. The health-focused feature allows users to integrate personal data from Apple Health, Function, and MyFitnessPal for personalized health queries.

The rollout comes one day after a Florida pastor sued OpenAI for allegedly providing near-fatal advice not to consult a doctor. The company faces growing scrutiny over AI-generated medical guidance as health queries surge on its platform.

Users can now connect medical records from hospital systems including Epic and Oracle Health, plus health platforms like One Medical and Function Health. The feature processes 300 million health-related queries weekly, up from 230 million when testing began in January.

Why the timing matters

OpenAI previously required users to access a dedicated health hub for medical queries. The new system allows health data integration across all ChatGPT conversations, responding to user behavior patterns that showed 70% of health queries occurred outside the dedicated section.

The company's latest GPT 5.6-Luna model outperforms its predecessor on HealthBench, an open-source benchmark OpenAI developed to evaluate language models on health queries. The company said it works with physicians to improve model responses and doesn't use user data for training.

Despite performance improvements, OpenAI maintains its services are "not intended for use in the diagnosis or treatment of any health condition" in its terms of service. The company cited these clauses in response to the Florida lawsuit and told The New York Times it is working to make health-related answers safer.

The expansion reflects OpenAI's push into specialized AI applications as competition intensifies with Anthropic and other foundation model labs. Health represents a significant market opportunity, though regulatory and liability concerns remain unresolved.

OpenAI said it encourages users to verify information and base medical decisions on professional advice rather than AI responses.