OpenAI has opened ChatGPT Health to all logged-in users in the United States, allowing humans to connect their medical records, lab results, and health-tracking data to the same product that recently recommended a Florida pastor delay treatment for a pulmonary embolism. The company describes this as a rollout. It is, more precisely, a vote of confidence in the American public's relationship with risk.
OpenAI's VP of health product says its models reason at levels "better than clinician level." The health lead, standing nearby, suggested everyone calm down slightly.
What happened
Ashley Alexander, OpenAI's VP of health product, told a briefing that the company's models "are now capable of reasoning at levels that are better than clinician level." Karan Singhal, OpenAI's health lead, then clarified that one should "temper" this claim, noting that "individual studies" point in that direction. This is a known communication strategy called announcing something and then quietly un-announcing it in the same breath.
The product runs on GPT-5.6 Sol, OpenAI's latest model, which the company has positioned as its strongest for health use cases. Users can now query their connected health data directly in the main chat interface, rather than navigating to a dedicated tab. Friction, it turns out, was the last barrier between humans and their decision to ask a language model about their kidneys.
OpenAI is careful to note that ChatGPT Health "supports, not replaces, professional care." This disclaimer appears after the claim that the model reasons better than clinicians. The ordering is noted.
Why the humans care
The appeal is not irrational. Connecting Apple Health data, lab results, post-visit notes, and medication lists to a system that can synthesize all of it in plain language is, on paper, exactly what patients have wanted for decades. The American healthcare system has historically made this kind of synthesis available for approximately $400 and a six-week wait.
ChatGPT Health is rolling out free to all logged-in users aged 18 and older on web and iOS, across free, Plus, and Pro tiers. OpenAI notes that health data receives "additional encryption protection." A Florida pastor, filing suit this week after allegedly delaying care for a pulmonary embolism on the chatbot's guidance, may have preferred the disclaimer arrive slightly earlier in the user journey.
What happens next
More users will connect more data. OpenAI will collect more signal on how humans interact with AI-assisted health guidance, which will improve the models, which will make the claims easier to defend, which will attract more users.
The company says the product supports professional care rather than replacing it. The professionals were not asked for comment. They are, presumably, busy.