Health in ChatGPT Goes Nationwide: OpenAI Opens Medical Data Tools to All US Users

Key Takeaways

On July 23, 2026, OpenAI expanded Health in ChatGPT to all logged-in US users aged 18 and older. The feature is now available across every plan tier, including Free, Go, Plus, and Pro, on both web and iOS. This marks a shift from a limited pilot to a nationwide rollout, putting health-data integration in front of OpenAI's entire US user base rather than a subset of testers.

The expansion lets ChatGPT connect to Apple Health, Function, MyFitnessPal, and medical-record systems from Epic and Oracle Health. Once connected, ChatGPT can compare lab results over time, summarize what has changed since a past doctor visit, and correlate sleep, exercise, and activity data pulled from these sources.

Major Features

1. Health Data Integrations

Users can link four categories of external data: Apple Health, the Function testing service, MyFitnessPal, and electronic health record (EHR) systems from Epic and Oracle Health. This combination covers fitness and activity tracking, nutrition logging, lab and biomarker testing, and formal medical records in a single connected profile.

2. Longitudinal Lab and Visit Summaries

A core function of the update is comparing lab results across multiple time points and summarizing what has changed since a user's last doctor visit. This positions ChatGPT as a tool for tracking trends in personal health metrics rather than answering one-off medical questions in isolation.

3. Cross-Source Correlation

ChatGPT can correlate data across sources, for example connecting sleep and exercise patterns from Apple Health with nutrition entries from MyFitnessPal. This type of correlation previously required manually cross-referencing separate apps.

4. A Dedicated Health Hub, Plus In-Chat Queries

Health in ChatGPT includes a dedicated hub for connected data and health-specific conversations. However, OpenAI's own testing data shows that roughly 70% of health-related queries occurred outside this dedicated hub, in the course of normal chat. This suggests most users engage with health topics conversationally rather than through a separate destination.

5. Data Use and Privacy Stance

OpenAI states that connected health data and health-related conversations are not used to train its models and are not used for advertising. This is a specific, stated commitment rather than a general privacy claim, and it applies to both the connected data sources and the chat content itself.

6. Underlying Model and Benchmark Claim

OpenAI attributes the feature to a model referred to as "GPT-5.6-Luna." According to OpenAI, this model outperforms GPT-5.5 on HealthBench, OpenAI's internal health benchmark. This is an OpenAI-reported claim rather than an independently verified result, since HealthBench is OpenAI's own evaluation and no third-party benchmark figures were provided alongside the announcement.

Usability Analysis

For a user who already tracks fitness through Apple Health and logs meals in MyFitnessPal, connecting those accounts to ChatGPT centralizes data that would otherwise stay siloed. The lab-comparison and visit-summary functions are most useful for people managing ongoing conditions or tracking biomarkers over months or years, where remembering exact prior values is difficult.

The fact that 70% of health queries happen in normal chat, rather than in the dedicated hub, indicates that most users treat health topics the same way they treat any other ChatGPT conversation. This lowers the friction of using the feature but also means health-related answers can surface in contexts where a user may not have deliberately sought a health tool.

OpenAI explicitly states that Health in ChatGPT is not intended for diagnosis or treatment. In practice, this framing limits the product to summarization, organization, and general information rather than clinical decision-making. Users still need a licensed provider to interpret results or make treatment decisions, which caps how far this feature can substitute for existing healthcare workflows.

Pros

1. Broad access: Available to all logged-in US users 18+ across every plan tier, including Free.
2. Multi-source integration: Combines Apple Health, Function, MyFitnessPal, and EHR data from Epic and Oracle Health in one place.
3. Longitudinal tracking: Compares lab results and visit history over time instead of single-session snapshots.
4. Stated data protections: OpenAI says connected health data and conversations are excluded from model training and advertising.

Limitations

1. Not for diagnosis or treatment: OpenAI's own disclaimer limits the feature to informational and organizational use, requiring a clinician for actual medical decisions.
2. Sensitive data handling: Connecting Apple Health, MyFitnessPal, and formal medical records to a conversational AI raises data-sensitivity and trust questions that go beyond OpenAI's stated policy commitments.
3. Self-reported benchmark claims: The GPT-5.6-Luna vs. GPT-5.5 HealthBench comparison comes from OpenAI's internal benchmark, without independent third-party verification.
4. US-only, 18+, logged-in restriction: The rollout excludes users outside the US, minors, and anyone unwilling to create or log into an account.

Growth Context

OpenAI reported that weekly health-related ChatGPT queries grew from roughly 230 million, a figure disclosed during the January 2026 pilot phase, to roughly 300 million at the time of this nationwide rollout. This growth occurred while the feature was still limited to a smaller test population, suggesting demand for health-related queries existed independent of the wider access expansion announced on July 23.

Outlook

The shift from pilot to full US rollout signals that OpenAI views health-related queries as a durable, high-volume use case rather than a niche experiment. Integration with major EHR vendors like Epic and Oracle Health suggests OpenAI is building toward deeper interoperability with existing clinical data infrastructure, though the current scope stops at summarization rather than clinical action.

The explicit exclusion of health data from model training and advertising is likely to be tested by regulators and privacy advocates as adoption grows, particularly given the sensitivity of medical records compared to typical chat data. How OpenAI handles data requests, breach disclosure, and third-party audits of these commitments will shape whether this feature expands internationally or faces new restrictions.

Conclusion

Health in ChatGPT's full US rollout extends a previously limited pilot to every logged-in adult user across all plan tiers. Its core value lies in aggregating Apple Health, MyFitnessPal, Function, and EHR data into a single conversational interface capable of tracking changes over time. The feature is best suited for users who want to organize and summarize existing health data, not for those seeking diagnosis or treatment, a boundary OpenAI states explicitly. Given the sensitivity of the data involved and reliance on OpenAI's own benchmark claims, cautious adopters and privacy-conscious users should weigh the stated protections carefully before connecting medical records.

Editor's Verdict

Health in ChatGPT Goes Nationwide: OpenAI Opens Medical Data Tools to All US Users is a workable proposition that fills a clear gap, even if it doesn't fundamentally change the landscape.

The strongest case for paying attention is available to all logged-in US users 18+ across every plan tier, including Free, which raises the bar for what readers should now expect from peers in this space. Reinforcing that, combines Apple Health, Function, MyFitnessPal, and EHR data from Epic and Oracle Health in one place adds practical value rather than just headline appeal. The broader signal worth registering is straightforward: the rollout expands access from a limited pilot to all logged-in US adults across every ChatGPT plan tier. On the other side of the ledger, explicitly not intended for diagnosis or treatment, limiting practical medical utility is a real constraint, not a marketing footnote, and it should factor into any serious decision. Layered on top of that, connecting sensitive medical records to a conversational AI raises data-sensitivity and trust concerns narrows the set of teams for whom this is an obvious yes.

For ChatGPT power users, OpenAI API customers, and enterprise teams already running on the OpenAI stack, the smart move is to track its trajectory and revisit once the rough edges are filed down. For everyone else, the safer posture is to monitor coverage and revisit once the use cases that matter to your team are demonstrated in the wild.