OpenAI’s ChatGPT Health rollout is expanding to all users in the United States on Thursday, giving more people the option to connect medical records and health-tracking data to ChatGPT, according to The Verge. The launch puts one of the most sensitive categories of personal data into OpenAI’s consumer chat interface, at the same time the company is making unusually bold claims about how its models perform on health-related reasoning.
During a briefing, Ashley Alexander, OpenAI’s vice president of health product, said the company’s models “are now capable of reasoning at levels that are better than clinician level,” The Verge reported. That is the kind of sentence that should make lawyers, doctors, and anyone with a pulse reach for the footnotes.
Asked how OpenAI’s models compare with human clinicians, OpenAI health lead Karan Singhal softened the claim. He said he would “temper” the statement that the models reason better than clinicians, while pointing to individual studies that have suggested movement in that direction, including research from Harvard and Stanford cited by The Verge.
What does ChatGPT Health do?
ChatGPT Health lets users connect health information to the chatbot and ask questions about it. OpenAI says the connected data can include post-visit notes, notes from a care team, lab results, and medications, and users can review or change which categories they have connected.
The new release changes where those questions happen. OpenAI will let users ask about connected medical information in the regular ChatGPT chat flow, rather than requiring them to go to a separate ChatGPT Health tab, according to The Verge.
OpenAI says ChatGPT will request permission before using a person’s health information to tailor an answer. The company also says conversations are encrypted at rest and in transit, while information connected through Health gets additional encryption protection. Those are security claims from OpenAI, not an independent audit of how the system behaves under pressure.
What model powers ChatGPT Health?
The wider release comes weeks after OpenAI introduced GPT-5.6 Sol, which the company has described as its “strongest model yet for health.” ChatGPT Health will run on OpenAI’s latest models, according to The Verge.
The mechanism is straightforward: users connect health records or tracker data, ChatGPT gets permission to use that information in a given exchange, and the model generates an answer in the main chat interface. The risk is just as straightforward. A fluent answer about symptoms, lab results, or medication can sound more authoritative than it is, especially when the system is reading personal records.
Why are OpenAI’s health claims under scrutiny?
OpenAI’s expansion lands as the company faces a lawsuit tied to alleged medical advice from ChatGPT. The New York Times reported Wednesday that a Florida pastor sued OpenAI, claiming ChatGPT gave him “extremely dangerous medical recommendations” and allegedly caused him to delay care for a pulmonary embolism.
That lawsuit is an allegation, not a finding. It still sharpens the central problem with health chatbots: OpenAI wants users to trust ChatGPT with medical context while also avoiding the implication that the product is a clinician.
OpenAI’s own executives appear to be walking that line. Alexander made a broad performance claim about model reasoning. Singhal then narrowed it by referring to individual studies rather than declaring a clean victory over doctors. For users, the practical takeaway is less glamorous than the product pitch: ChatGPT Health can organize and answer questions about connected health data, but OpenAI has not shown, in the facts reported so far, that it should replace medical care.
This story draws on original reporting from The Verge.
