OpenAI has announced that ChatGPT Health is now available to all U.S. users aged 18 and older, across every plan. That includes free, Go, Plus, and Pro accounts. The rollout is happening this week on the web and iOS for logged-in users.
The timing is awkward, to put it mildly. The announcement comes just one day after a Florida-based pastor filed a lawsuit against OpenAI, claiming the chatbot gave him a suggestion not to see a doctor, advice he described as nearly fatal. OpenAI cited its own terms of service in response to the suit, noting that its services are “not intended for use in the diagnosis or treatment of any health condition.”
That disclaimer did not stop the company from pushing the feature to millions more people. OpenAI says it wants users to verify information and base medical decisions on advice from actual professionals, but it is clearly betting that the demand is there regardless.
ChatGPT Health started as a limited test in January, with a dedicated hub where users could connect data from services like Apple Health, MyFitnessPal, and Function. The feature also supports medical records from major hospital systems, including Epic and Oracle Health, and platforms like One Medical and Function Health. The idea is to let ChatGPT pull in personal health context when answering questions.
One of the bigger changes in this rollout is that the health data is no longer siloed inside the dedicated hub. Users can now draw on their connected health information in any conversation. That shift came after OpenAI noticed that 70% of health-related queries were already happening outside the hub anyway. So instead of pushing users to a separate section, the company is bringing the context to wherever the conversation is happening. Someone asking about a restaurant menu, for example, could get a response that accounts for their recorded allergies.
The numbers behind this feature are significant. When OpenAI first launched the health hub in January, users were sending 230 million health-related queries per week. That figure has since climbed to 300 million. That kind of volume explains why the company is investing here, even with the legal and ethical risks attached.
On the technical side, OpenAI says its models are getting better at handling health questions. The company’s smallest model from its latest release, GPT 5.6-Luna, outperforms GPT 5.5 on HealthBench, an open-source benchmark OpenAI built to test how well large language models handle health queries. OpenAI also says it works with physicians to improve its models and does not use personal health data to train them.
Still, the broader picture here is complicated. Multiple studies have found that AI chatbots are not reliable sources of medical advice. Accuracy varies widely depending on the question, the model, and how the information is framed. OpenAI is not alone in moving forward anyway. Anthropic and Google have both launched health-focused AI features in recent months, and none of them have let the research warnings slow them down much.
The core tension for OpenAI, and for the industry at large, is this: people are clearly using these tools for health questions whether companies build dedicated features or not. Ignoring that behavior does not make it safer. Building guardrails around it might. But a lawsuit filed the day before a major health rollout is a reminder that the guardrails are still very much a work in progress.
