OpenAI began rolling out Health in ChatGPT to all US users over 18 on July 23, 2026, one day after a Florida pastor sued the company over a near-fatal medical suggestion.
Quick Summary – TLDR:
- OpenAI opened Health in ChatGPT to every US user over 18 across the Free, Go, Plus, and Pro plans on web and iOS.
- ChatGPT now fields more than 300 million health-related queries each week, up from 230 million during a January test.
- Apple Health, Epic and Oracle Health records, One Medical, and Function Health can now feed a user’s ChatGPT conversations.
- A Florida pastor sued OpenAI the day before launch, alleging ChatGPT told him not to see a doctor.
- OpenAI says connected health data is never used to train its foundation models or to target ads.
What Happened?
OpenAI made Health in ChatGPT available to every logged-in US user aged 18 and older, the company said in a newsroom post. The feature spans the Free, Go, Plus, and Pro tiers on web and iOS, though it is not yet available in Codex.
The launch ends a staged test that began in January, when health tools lived in a separate hub inside ChatGPT. OpenAI found that 70% of health-related conversations happened outside that hub, so it now lets users pull connected health data into any chat. People can also type @Health to bring that context in on demand.
Health in ChatGPT is starting to roll out to U.S. users.
— ChatGPT (@ChatGPTapp) July 23, 2026
Open Health from the sidebar to securely connect and manage accounts, view a timeline of synced records, and explore recent data and trends.
With your permission, ChatGPT can draw on your connected medical records and… pic.twitter.com/nCNZfWAWlw
The lawsuit shadowing the launch
The wider release landed one day after a Florida-based pastor sued OpenAI, alleging that ChatGPT gave him a near-fatal suggestion not to consult a doctor, according to The New York Times. The suit is an allegation that has not been tested in court.
OpenAI has pointed to its terms, which state the service is “not intended for use in the diagnosis or treatment of any health condition.” The company said it is working to make health and medicine answers safer, and it urges users to verify information and base medical decisions on professional advice. Studies flagged by TechCrunch, including one from the University of Oxford, have separately warned that popular AI chatbots often give problematic or unreliable medical advice.
Inside the models behind the answers
OpenAI ties the launch to model gains it says improve safety. GPT-5.5 Instant, which serves free users, made progress at recognizing when urgent care is needed and explaining uncertainty, the company said. GPT-5.6 Sol, offered to paid users, is positioned as its strongest health model yet.
OpenAI also says GPT-5.6 Luna, the smallest model in its latest release, outperforms GPT-5.5 on HealthBench Professional, an open-source benchmark the company built to score AI on health tasks. Every GPT-5.6 model beat GPT-5.5 on that evaluation, which hundreds of physicians helped design, according to the company.
Privacy controls and what users can connect
Users can link Apple Health, medical records from US hospital systems such as Epic and Oracle Health, plus One Medical and Function Health. TechCrunch also lists MyFitnessPal among the connected apps. By default, ChatGPT asks permission before using connected data, and users can switch to always-allow in Settings, Plugins, Health.
Anyone who connects records should confirm the details are current, since a medication can stay listed after someone stops taking it, and should check important facts against the original source. Disconnecting an account deletes that synced data from OpenAI’s systems within 30 days, though information already in a chat stays until that chat is deleted. Using Temporary Chat, or turning memory off, helps reduce the chance that health details persist into future conversations, a control worth knowing as usage climbs.
SQ Magazine’s Takeaway
Making health context available in every chat removes the friction that kept most users out of the dedicated hub, and it arrives as OpenAI competes with Anthropic’s Claude for Healthcare and Google’s Fitbit health coach for the same consumer moment. The wager is that scattered records, wearables, and lab results become more useful when one assistant can read across them, with the company framing the tool as support for professional care rather than a replacement for it.
The unresolved question is whether safety keeps pace with reach. OpenAI expanded to the full US adult base the same week a lawsuit put its medical guidance in front of a court, and independent studies still flag unreliable chatbot health advice. What OpenAI has not detailed is how it measures real-world harm now that connected records raise the cost of a wrong answer, how many of its hundreds of millions of weekly queries involve urgent symptoms, or how the permission prompts hold up under red-team pressure at scale. Regulators and plaintiffs’ attorneys are likely to watch closely, and verifying anything that matters against a clinician remains the safe default.
