ChatGPT Health Explained: Who Gets It, Is It Free, and Is It Safe?

The short version: ChatGPT now has a dedicated Health space, and as of July 23, 2026 it is open to everyone in the US -- no waitlist. If you are a logged-in ChatGPT Free, Go, Plus, or Pro user aged 18 or older, on the web or the iOS app, you can securely connect Apple Health and your medical records and then ask ChatGPT questions grounded in your own numbers. It is US-only, entirely opt-in, and OpenAI is explicit that it is not a diagnosis tool. This page explains what it does, who actually gets it, and whether it is safe to hand it your medical records.

What Changed (July 23)

ChatGPT Health moved from a limited waitlist to open availability in the US. OpenAI first started testing Health with a small group of early users earlier in 2026; on July 23, 2026 it added Health to the ChatGPT release notes as broadly available, stating that logged-in Free, Go, Plus, and Pro users in the United States who are 18 or older can now connect their health data. In plain terms: if you are in the US, you no longer have to request access. OpenAI does add one caveat that a lot of coverage dropped, in its help article: "We're rolling it out gradually, so some eligible users may not see it right away." On iOS you also need ChatGPT version 1.2026.188 or later. If the Health tab is not in your sidebar, update the app before concluding you are not eligible.

What Changed (September 28)

Your Health dashboard became something you can ask questions of directly. OpenAI's release notes entry of September 28, 2026 says you "can now select a chart, metric, or record in the Health tab to get personalized explanations and insights based on your connected health information." Before this, the way to ask about a number was to describe it in chat; now you tap the number itself and ChatGPT explains it against the rest of what you have connected. OpenAI published no plan, country or age qualifier with that entry, so the existing gates below are the ones to assume: US, 18 or older, on every plan.

Key Takeaways

• It is live in the US now. OpenAI opened Health in ChatGPT to all logged-in Free, Go, Plus, and Pro users in the United States, aged 18 or older, on July 23, 2026.
• It is free on every plan. Health is not a paid upgrade -- it is included whether you pay for ChatGPT or not. What changes across plans is your usual message limits and the model answering: Free and Go now default to GPT-5.6 Luna, per OpenAI's GPT-5.6 help page, while GPT-5.6 Sol, "available to paid users, is our strongest model yet for health".
• US and 18+ only. You must be in the United States and at least 18 to connect medical records. There is no Android app or non-US access yet.
• You connect the data, on purpose. Health only knows what you choose to link. OpenAI names four sources: Apple Health, supported medical records from US hospital systems, One Medical, and Function Health. Nothing connects automatically.
• Your data is walled off. OpenAI says connected health data and the conversations that use it are not used to train its models or target ads, and are encrypted.
• It is not a diagnosis tool. OpenAI states Health is meant to support, not replace, medical care, and is not intended for diagnosis or treatment.
• The default changed on September 14, 2026, and it went the permissive way. OpenAI says that for new Health connections, ChatGPT "will now default to each user's global plugins permissions setting," and that if you have not changed yours, that default is "Allow low-risk actions." In plain terms: it can use your connected health data when relevant without asking each time. Sensitive actions, such as emailing your data to someone, still prompt. Change it at Settings > Plugins > Health.
• You can now tap a chart or a number and ask about it. Since September 28, 2026 OpenAI says selecting "a chart, metric, or record in the Health tab" produces a personalized explanation drawn from your connected health data.

What Is ChatGPT Health?

ChatGPT Health is a dedicated space inside ChatGPT for health and wellness questions, where you can connect your own health data so answers are grounded in your actual records instead of generic advice. Once you link Apple Health and supported medical records, ChatGPT can see information like lab results, medications, visits, sleep, and activity, and use it when you ask a relevant question. OpenAI describes it as a place to view your health information in one dashboard and ask questions grounded in your personal context -- comparing a new lab result to past ones, summarizing what changed since your last appointment, or working out what to raise with a doctor.

The practical difference from just typing a health question into ChatGPT is memory of you. A normal chat knows nothing about your cholesterol history; Health can, if you connect the records, answer "is this number better or worse than last year" using your own results.

Do You Get It, and on Which Plan?

Yes -- if you are in the US and 18 or older -- and it does not depend on paying. Here is where things stand as of July 23, 2026:

The gate here is not your wallet -- it is your location and age. This is unusual: most new ChatGPT features arrive first for paid users. Health is the opposite, available on the Free plan from day one, but fenced to US adults on web and iOS. If you are outside the US, no plan unlocks it yet.

One qualifier on "the same on every plan", added after re-reading OpenAI's announcement in September 2026: the feature is the same, but the model behind it is not. OpenAI's announcement says GPT-5.5 Instant, "available to users on the Free plan", made "meaningful gains" on health questions, while GPT-5.6 Sol, "available to paid users, is our strongest model yet for health". The Free half of that is out of date: OpenAI's GPT-5.6 help page says the Free and Go "default model is GPT-5.6 Luna," and GPT-5.5 retired on October 14, 2026, in a retirement OpenAI scoped to Codex with ChatGPT sign-in rather than to chat. So free access is real and complete, and paying still buys better health answers, not just more of them.

What Can It Actually Do?

It turns your scattered health records into something you can ask questions about in plain English. After you connect data, according to OpenAI's own description, you can:

• Understand test results -- ask what a lab value means and how it compares to your previous results.
• Prepare for appointments -- summarize what has changed since your last visit and draft questions to ask your doctor.
• Track trends -- look at sleep, activity, and other metrics over time in one place rather than across separate apps.
• Think through concerns -- talk through symptoms or worries to decide what is worth raising with a professional.

What it will not do is diagnose you or tell you what treatment to take. OpenAI is direct that Health "is designed to support, not replace, medical care and is not intended for diagnosis or treatment." Treat it as a smart way to read and organize your own records before a real appointment, not as a substitute for one.

Is It Safe to Connect Your Medical Records?

This is the real question, and OpenAI's answer is more reassuring than most AI features. Per OpenAI, connected medical records and Apple Health data, and the conversations that use them, are not used to train its foundation models or to target ads. Your health data is encrypted, and Health is built with what OpenAI calls layered privacy and security safeguards that keep these conversations separate from the rest of ChatGPT.

You still control the switch, but read the default carefully, because OpenAI changed it. Until September 14, 2026 a new Health connection asked permission each time before using your records. OpenAI's release note of that date says new Health connections now "default to each user's global plugins permissions setting," and "if you make no changes to your Plugins settings, the default permission setting is Allow low-risk actions" -- so ChatGPT "can use it when relevant without asking for permission each time, unless you are attempting a sensitive action, such as asking ChatGPT to email a training plan based on your data to a running partner." OpenAI's stated reason is usage: "more than 70% of Health users already choose to allow ChatGPT to use data connected in Health without asking for permission each time." You choose which accounts to connect, you can turn on always-allow if the prompts get tedious (the switch is at Settings > Plugins > Health), and if you disconnect an account the connected data is deleted from OpenAI's systems within 30 days -- disconnect at Health > Accounts. Going the other way, you can type @Health in any message to tell ChatGPT to use your health context deliberately rather than waiting to be asked. One thing worth knowing: memories can be created from your health conversations (not directly from the connected records), so if you want a private one-off, use Temporary Chat or turn memory off.

Should You Turn It On?

Turn it on if you are in the US, already use ChatGPT, and juggle real health data -- lab results, a chronic condition, multiple apps. For those people, having ChatGPT read your own numbers and prep you for appointments is a genuine time-saver, and the privacy terms (no training, no ads, delete-on-disconnect) are about as good as this gets. Skip it if you are outside the US (you cannot use it anyway), you are uneasy handing any AI your medical records, or you would be tempted to treat its answers as a diagnosis -- that is exactly what OpenAI says not to do.

ChatGPT Health is part of a wider push to make ChatGPT the app you keep open all day -- see our roundup of the July 2026 ChatGPT app changes and, if you are weighing whether to pay at all, which AI chatbot is actually worth paying for. If you want the model side of the story, GPT-5.6 Sol, explained covers what powers ChatGPT today.

The bottom line: ChatGPT Health is free, live for US adults on web and iOS, and privacy terms are unusually strong for an AI feature -- but it is US-only and firmly not a diagnosis tool. If you are in the US and already lean on ChatGPT, connecting your records is a low-risk way to make it genuinely useful for your health. Everyone else is waiting on a rollout OpenAI has not dated.

Corrections and Updates

September 17, 2026: this page said ChatGPT asks permission each time before using your medical records. That is no longer the default. OpenAI's release note dated September 14, 2026 says new Health plugin connections now default to your global plugins permission, which is "Allow low-risk actions" unless you changed it, so ChatGPT "can use it when relevant without asking for permission each time." Sensitive actions still prompt. The Key Takeaway and the privacy section were corrected, and the setting path (Settings > Plugins > Health) is unchanged.

September 16, 2026: this page said GPT-5.5 Instant serves the Free plan; Free and Go now default to GPT-5.6 Luna. The line repeated OpenAI's Health announcement, which predates the change on OpenAI's GPT-5.6 help page. GPT-5.5 retires from ChatGPT on October 14, 2026. (Superseded October 14, 2026: OpenAI scoped that retirement to Codex with ChatGPT sign-in; it has described no change to ChatGPT chat.)

September 2, 2026: re-verified, with three additions and one correction.