On July 23, 2026 (U.S. time), OpenAI began rolling out Health, a new feature in ChatGPT for working with your health information, to users in the United States[1]. Once you securely connect Apple Health and supported medical records, you can use your own data to understand test results, prepare for appointments, and rethink daily habits through more informed conversations. The feature is available across all plans, including the free tier, starting on the web and iOS.
Bringing scattered health information together in ChatGPT
According to OpenAI, more than 300 million people turn to ChatGPT with health-related questions every week[1]. The problem is that information such as lab results, medication history, visit records, and wearable data tends to be scattered across patient portals and different apps, making it hard to see the full picture.
Health lets ChatGPT draw on the information you have connected, within the limits you allow. You can compare new results against prior tests, summarize what has changed since your last visit, or explore how sleep, activity, and workouts relate to your daily routine[1]. The goal is to cut down on repeatedly explaining the same details or re-uploading documents, and to help you ask sharper questions.
Supported connections include Apple Health, medical records from participating U.S. hospital systems, One Medical, and Function Health[1]. Health, in the left sidebar, serves as the hub for managing your information, checking recent data and trends, browsing synced records, and revisiting past health conversations. You can also still ask health questions without connecting anything at all.
Using health information in the flow of conversation
A defining part of the design is that you can draw on connected information during ordinary conversations, without opening a dedicated screen. OpenAI says that earlier this year it offered a dedicated health space to a limited group of users, but when it looked at usage, more than 70% of health-related conversations happened outside that dedicated space[1]. Health topics naturally come up in everyday exchanges, such as planning meals or looking for recipes.
So with the new Health, ChatGPT can reference the information you have allowed across your conversations, without starting inside Health. For example, it can take a dietary restriction into account when helping you pick a restaurant, or a recent injury when suggesting lower-impact weekend plans[1]. Adding @Health to a message lets you explicitly ask ChatGPT to include your health context. The feature itself is a broader, official relaunch of the health experience that was available to a limited group in the first half of 2026[2].
GPT-5.5 Instant and GPT-5.6 Sol, tuned for health
Behind Health are models given additional training for the health domain. GPT-5.5 Instant, available on the free plan, is said to have improved at recognizing when urgent care may be needed, asking for missing context, and explaining uncertainty[1]. GPT-5.6 Sol, available to paid users, is positioned as OpenAI's strongest model yet for health, with greater ability to reason across complex details.
To measure performance, OpenAI works with hundreds of physicians around the world to build realistic health scenarios and evaluation rubrics. These assess qualities such as accuracy, safety, clarity, and appropriate escalation (handing off to a professional). On HealthBench Professional, which measures performance on challenging health tasks, every GPT-5.6 model outperformed GPT-5.5[1].
How privacy and security are handled
Because it deals with sensitive health data, OpenAI says it layered in multiple safeguards. Connected medical records, Apple Health information, and conversations that use them are explicitly not used to train foundation models or to target ads[1]. Conversations are encrypted at rest and in transit, and information connected in Health receives additional encryption protections.
Control stays with the user. By default, ChatGPT asks for permission before using connected information in a response, and you can approve each request or choose to always allow access[1]. You can disconnect at any time, and when you do, data synced from that source is deleted from OpenAI's systems within 30 days.
Points to keep in mind
At the same time, availability is currently limited to logged-in users aged 18 and older in the U.S., and Health is not yet available in Codex, OpenAI's coding tool[1]. OpenAI itself repeatedly stresses that ChatGPT can make mistakes and does not replace the care and judgment of medical professionals, urging people to verify important information against original sources and to discuss medical decisions with their doctor.
There is also understandable caution about entrusting sensitive medical data to a consumer AI service. OpenAI responds with non-training policies, additional encryption, and permission controls, but how far people will feel comfortable relying on it is something that will be tested in real-world use.
Summary
Health consolidates health information once scattered across patient portals and apps into ChatGPT, letting you have conversations grounded in your own situation within the natural flow of chatting. It has launched in the U.S., including on the free plan, built on health-tuned models such as GPT-5.6 Sol. Foregrounding privacy protections while positioning the tool as a complement to, not a replacement for, professional care is the crux of this announcement.
