OpenAI has introduced Health in ChatGPT, a new feature that allows users to securely connect Apple Health and supported medical records so ChatGPT can use relevant health information during conversations. Users can choose what information they connect and control when ChatGPT can access it.
Every week, more than 300 million people use ChatGPT for health-related questions, including understanding lab results, preparing for appointments, interpreting information shared by doctors, and building healthier routines. OpenAI said health information is often spread across patient portals, medical records, apps, and wearable devices, making it difficult to view complete health context.
With permission, ChatGPT can use connected Health information to help users compare new results with previous tests, summarize changes since their last appointment, review medications, and understand how sleep, activity, and workouts relate to their routine.
Health conversations across ChatGPT
OpenAI previously introduced a dedicated Health experience to a limited group of users. Early testing showed that health questions often appeared naturally during regular conversations, with more than 70% of health-related conversations among users with access taking place outside the dedicated Health experience.
With the updated experience, users can allow ChatGPT to consider connected information such as:
- Medications
- Lab results
- Recent visits
- Sleep data
- Activity information
This information can be used alongside the goals and context shared by users. For example, ChatGPT can consider dietary restrictions when helping with food choices or take a recent injury into account when suggesting lower-impact activities.
Users can ask follow-up questions, add details, or correct information that may be incomplete or outdated.
The Health section in ChatGPT allows users to:
- Connect and manage health accounts
- View recent data and trends
- Browse synced records
- Explore suggested prompts
- Access previous health conversations
- Review and update conditions and medications
- Add missing details such as family health history
OpenAI said connected information may not always be complete or current. For example, a medication may remain listed after a user stops taking it. Users can update information when changes occur and verify important details with original sources.
Health improvements with GPT-5.5 Instant and GPT-5.6 Sol
OpenAI said it continues improving the models used for health conversations in ChatGPT.
GPT-5.5 Instant, available to Free users, improves the ability to recognize when urgent care may be needed, ask for relevant context, explain uncertainty, and simplify complex health information. On OpenAI’s most challenging health evaluations, GPT-5.5 Instant performed at a level comparable to its frontier Thinking models at the time of evaluation.
GPT-5.6 Sol, available to paid users, improves performance on tasks requiring reasoning across complex health details, clear communication, and careful judgment. It can help explain medical visit notes, understand changes in lab results over time, consider factors involved in options discussed by doctors, and prepare questions for follow-up appointments.
OpenAI works with hundreds of physicians worldwide to develop health scenarios and evaluation criteria. These evaluations measure:
- Accuracy
- Safety
- Communication
- Context awareness
- Completeness
- Appropriate escalation
OpenAI said GPT-5.6 models outperformed GPT-5.5 on HealthBench Professional, an evaluation focused on challenging health tasks.
Physicians also tested Health in ChatGPT before release to evaluate model performance and safety when using connected health information. OpenAI said ChatGPT can still make mistakes and does not replace the care and judgment of qualified medical professionals.
Privacy, security, and user controls
OpenAI said Health in ChatGPT includes privacy and security safeguards. ChatGPT conversations are encrypted at rest and in transit, while information connected through Health receives additional encryption protections.
Connected medical records, Apple Health information, and conversations that use this data are not used to train OpenAI’s foundation models or target advertisements, regardless of the model training setting selected by users.
Users control when ChatGPT can access connected Health information:
- ChatGPT requests permission before using connected medical records and Apple Health information by default.
- Users can approve individual requests or enable continuous access.
- These settings can be changed through Settings > Plugins > Health.
Conversations that do not use Health information continue to follow standard ChatGPT model-training settings, which can be managed through Settings > Data Controls.
Users can disconnect Health accounts anytime through Health > Accounts. When disconnected, synced data from that source is deleted from OpenAI systems within 30 days. Information already included in ChatGPT conversation history remains until users delete those conversations.
Memory settings remain under user control. Memories can be created from Health conversations, but memories are not created directly from connected medical records or Apple Health information. Users can use Temporary Chat or disable memory to prevent conversations from creating memories.
Connecting and using Health information
Users can continue asking health-related questions in ChatGPT without connecting any accounts. To use responses based on personal health information, users can connect:
- Apple Health
- Supported medical records from U.S. hospital systems
- One Medical
- Function Health
To connect health information, users can open Health from the ChatGPT sidebar or More menu and select “Get started.” Once connected, information begins syncing, which may take a few minutes.
Users with wearable devices, fitness apps, or nutrition apps connected through Apple Health can allow ChatGPT to access the information shared through Apple Health. Available information depends on the connected app and supported metrics, and some proprietary scores may not transfer.
After syncing, users can ask questions anywhere in ChatGPT. When relevant, ChatGPT may request permission before using Health information. Users can also add @Health to a message to explicitly request Health context.
Availability
Health in ChatGPT is rolling out to logged-in users aged 18 and older in the U.S. across Free, Go, Plus, and Pro plans. The feature is available on web and iOS and is not currently available in Codex.