OpenAI integrates medical records into ChatGPT for real-time health insights

The redesign marks a shift from a siloed approach, where users previously had to enter a dedicated health mode to access their data. OpenAI’s internal research found that over 70 percent of health-related queries occurred during unrelated tasks, such as meal planning or symptom checking. By integrating the data directly into the chat flow, the system can now offer personalized advice—such as suggesting dietary adjustments based on logged restrictions or modifying fitness goals based on recent injury reports.

To power these features, OpenAI introduced two specialized models: GPT-5.5 Instant for free users and the more robust GPT-5.6 Sol for paid tiers. The company claims these models were vetted by hundreds of physicians using "HealthBench Professional" rubrics to ensure accuracy and safe escalation protocols. While early testers, including patients and medical professionals, reported success in synthesizing complex medical timelines, the company maintains that the tool is strictly for support, not clinical diagnosis.

Privacy remains a core concern, and OpenAI has instituted a permission-based architecture. Connected health data is excluded from foundation model training and ad targeting, regardless of a user’s global privacy settings. However, the system faces potential friction regarding data quality; because synced records may become stale or incomplete, the company advises users to manually flag changes and verify critical information with their healthcare providers. Access is currently limited to users aged 18 and older in the United States.