Health in ChatGPT Launch

OpenAI has launched Health in ChatGPT for users in the United States, enabling the AI to securely integrate with Apple Health and supported medical records. This integration allows ChatGPT to provide personalized health insights by analyzing a user's medical history, lab results, and activity data, reducing the need for users to manually upload or explain their health context.

Personalized Health Context Across Conversations

ChatGPT can now draw on connected health data—including medications, lab results, recent visits, sleep, and activity—to personalize responses across all conversations, not just within a dedicated health space. This shift follows early testing where over 70% of health-related queries occurred naturally during general interactions, such as meal planning or activity scheduling.

Users can manage their health data through a dedicated Health section in the sidebar, which serves as a hub for connecting accounts, tracking trends, and browsing synced records. To explicitly trigger health context, users can use the @Health command in their messages.

Model Improvements and Health Intelligence

OpenAI has updated its models to improve reasoning and communication in health-related contexts. These improvements are driven by dedicated training and collaboration with hundreds of physicians worldwide to develop realistic scenarios and evaluation rubrics.

GPT-5.5 Instant

Available to all free users, GPT-5.5 Instant focuses on recognizing when urgent care is needed, explaining uncertainty, and simplifying complex medical information. During evaluations, it performed at a level comparable to frontier Thinking models.

GPT-5.6 Sol

Available to paid users, GPT-5.6 Sol is the strongest model for health tasks. It demonstrates superior performance on the HealthBench Professional evaluation, specifically in reasoning across complex details and providing careful judgment for tasks such as explaining visit notes or analyzing changes in lab results over time.

Privacy, Security, and User Control

Health in ChatGPT is built with layered safeguards to protect sensitive medical data. All conversations are encrypted at rest and in transit, with connected health information receiving additional encryption protections.

• No Model Training: Connected medical records, Apple Health information, and conversations that utilize this data are not used to train foundation models or target advertisements, regardless of the user's general model-training settings.
• Explicit Permission: By default, ChatGPT asks for permission before accessing connected health data to personalize a response. Users can choose to allow access on a per-request basis or enable "always allow" in settings.
• Data Deletion: Users can disconnect health accounts at any time. Data synced from those sources is deleted from OpenAI's systems within 30 days.
• Memory Management: Memories created from health conversations are not derived directly from connected medical records or Apple Health data. Users can use Temporary Chat or disable memory entirely to prevent the history from being saved.

Integration and Availability

Health is currently rolling out to logged-in users aged 18 and older in the U.S. across Free, Go, Plus, and Pro plans on web and iOS. It is not yet available in Codex.

Users can connect Apple Health, One Medical, Function Health, or supported medical records from U.S. hospital systems. For those using third-party wearables or nutrition apps, data is synced if the app shares information with Apple Health.

User Impact

Early testers have highlighted the ability to synthesize fragmented medical histories into usable timelines. As noted by Blake, a technical program manager:

"The most useful part has been turning scattered medical history into something I can actually understand and explain... I could understand the bigger pattern. It made the information more usable and gave me better language to advocate for myself with providers and trainers."

OpenAI emphasizes that ChatGPT is a tool to support, not replace, the care and judgment of qualified medical professionals.