OpenAI's ChatGPT Health feature is now available to all users aged 18 and older in the United States, spanning Free, Go, Plus, and Pro subscription plans. The rollout comes just one day after a Florida pastor filed a lawsuit against the company, claiming that the chatbot's advice nearly cost him his life.

From Dedicated Hub to Every Conversation

Previously, users seeking health-related answers had to navigate to a dedicated hub — a pilot program OpenAI launched in January 2026. That restriction has now been lifted, allowing ChatGPT to draw on a user's connected health data during any conversation. For instance, if a user asks about food or allergies in a standard chat, the assistant can reference relevant information the user has linked to those topics.

OpenAI reported that its internal testing found approximately 70% of health-related activity was already occurring outside the hub. The company also disclosed that weekly health-related queries have grown from 230 million in January to 300 million currently. Logged-in users in the US on the web and iOS platforms will gain access this week as the feature rolls out. It is not yet available in OpenAI's coding tool, Codex.

Health in ChatGPT is starting to roll out to U.S. users. You can securely connect Apple Health and supported medical records to understand your information in context, track what has changed, and have more informed conversations. — OpenAI (@OpenAI) July 23, 2026

Connected Data and Model Improvements

The functionality depends on the data sources users choose to link. On iPhone, users can connect Apple Health along with services such as Function and MyFitnessPal. Medical records from hospital systems including Epic and Oracle Health can also be integrated, and platforms such as One Medical and Function Health are supported as well. Once connected, ChatGPT can display a new lab result alongside previous tests and highlight what has changed since the user's last interaction.

That shift puts ChatGPT closer to a personal health-information organizer than a general chatbot, because its answers can be shaped by records, wearable data, nutrition logs, and prior results supplied by the user. It also makes the boundaries of the product more important: connected data may help users ask more informed questions, but it does not replace a clinician's interpretation of symptoms, test results, or treatment decisions.

OpenAI stated that the underlying models have improved. GPT 5.6-Luna, the smallest model in the latest release, outperforms the previous GPT 5.5 on HealthBench — an open-source benchmark the company developed to evaluate how effectively language models handle health-related questions. The company also noted that it does not use customer data to train its models and collaborates with physicians to refine them.

Lawsuit Casts Shadow Over Launch

On the eve of the announcement, a Florida pastor sued OpenAI, alleging that the chatbot advised him not to consult a doctor — guidance he says nearly proved fatal.

OpenAI responded by citing its own terms of service, which state that the platform is "not intended for use in the diagnosis or treatment of any health condition." The company also told The New York Times that it is actively working to make its health and medical responses safer. With this launch, OpenAI emphasized that it expects users to fact-check information and base medical decisions on professional advice.

The timing underscores the tension facing consumer AI health tools as they move from isolated pilots into everyday conversations. For users, the most important details to watch are how clearly ChatGPT distinguishes informational guidance from medical advice, how consent and data connections are managed, and how consistently the system directs people to qualified professionals when health decisions require clinical judgment.