ChatGPT Health Is Now Available to Everyone – Here’s Everything You Need to Know

ChatGPT Health is OpenAI’s built-in feature for answering health and wellness questions using both general medical knowledge and, if you choose to connect them, your own personal health data — from Apple Health steps and sleep data to lab results synced from providers like Epic. As of this week, it’s available to every logged-in U.S. user over 18, on every ChatGPT plan, on web and iOS.

What Changed in This Rollout

From Test Hub to Every Chat

When OpenAI first introduced Health back in January, it lived inside a separate, dedicated hub — you had to go looking for it. That’s no longer the case. Health-related context can now surface inside any regular conversation, not just the special section built for it.

That shift matters more than it sounds. During the testing period, OpenAI found that roughly 70% of health-related questions were happening outside the dedicated hub anyway — people were asking about symptoms, food, or allergies in the middle of everyday chats, not in a separate “health mode.” Folding the feature into general chat simply matches how people were already using it.

Who Can Use It Now

The rollout covers logged-in U.S. users on Free, Go, Plus, and Pro plans, on both the web app and iOS, starting this week. It’s restricted to users 18 and older.

How ChatGPT Health Actually Works

Connecting Apple Health, MyFitnessPal, and Fitness Apps

Users can link personal data from consumer apps like Apple Health, MyFitnessPal, and Function to give ChatGPT context — think activity levels, nutrition logs, sleep patterns, or heart rate trends. Once connected, that data can inform answers even in unrelated conversations, like a question about which snack fits your macros for the day.

Linking Medical Records

Beyond fitness apps, ChatGPT Health also supports connecting actual medical records — pulling data from hospital and health-record systems such as Epic and Oracle Health, plus platforms like One Medical and Function Health. In theory, this lets the assistant reference things like recent lab work or diagnosis history when you ask a related question.

Step-by-Step: Turning On ChatGPT Health

1. Open ChatGPT on web or iOS and make sure you’re logged in with a U.S. account.
2. Go to the Health section in Settings (or the dedicated Health hub, if still shown on your account).
3. Choose which services to connect — Apple Health, MyFitnessPal, Epic, Oracle Health, One Medical, or Function Health.
4. Authorize each connection through that service’s standard permissions screen.
5. Start chatting normally — connected health context can now inform answers anywhere in ChatGPT, not just inside the Health hub.

You can disconnect any linked service at any time from the same settings menu if you change your mind.

The Numbers Behind the Launch

OpenAI’s own figures show how fast health-related usage has grown. Back when Health first launched in testing, the company said users were sending around 230 million health-related messages every week. That number has since climbed to 300 million a week — a roughly 30% increase in just a few months, and a clear signal of how much people already lean on ChatGPT for health questions, feature or no feature.

Is ChatGPT Health Accurate? What HealthBench Actually Measures

OpenAI evaluates its models against HealthBench, an open-source benchmark it built specifically to test how well large language models handle health-related questions, developed in collaboration with physicians. According to OpenAI, even its smallest newer model, GPT-5.6-Luna, scores higher on HealthBench than the previous GPT-5.5 model.

That’s a genuinely useful data point — but it’s worth reading carefully. A benchmark score tells you a model is improving at structured health-question tasks; it doesn’t mean the model is a substitute for a licensed clinician evaluating your specific case, medical history, and symptoms in person. OpenAI itself doesn’t claim otherwise: its own terms of service state its products are “not intended for use in the diagnosis or treatment of any health condition.”

Privacy: Does OpenAI Use Your Health Data to Train Its Models?

According to OpenAI, user data connected through Health is not used to train its models, and the company says it works with physicians to refine how its models handle health questions. If privacy is a concern, it’s worth reviewing exactly which data categories you’re authorizing before connecting any third-party health service — and remembering that you can revoke access at any time.

Why This Launch Is Controversial Right Now

The Lawsuit Behind the Timing

The timing here isn’t a coincidence. This nationwide rollout arrived just one day after a Florida-based pastor filed a lawsuit against OpenAI, alleging the chatbot gave a near-fatal suggestion that discouraged him from seeing a doctor. OpenAI has pointed to its terms of service — which explicitly state the product isn’t meant for diagnosis or treatment — in response to the suit, and has told reporters it’s working to make health-related answers safer going forward.

What Independent Research Says About AI Health Advice

The lawsuit isn’t an isolated data point. Multiple independent studies, including research from the University of Oxford and coverage summarized by News-Medical, have raised concerns that AI chatbots — across providers, not just OpenAI’s — can give problematic or unreliable medical advice in certain scenarios. None of this has stopped the broader industry from moving forward, though: Anthropic launched Claude for Healthcare in January, and Google added an AI-powered health and fitness coach to Fitbit devices in 2025.

ChatGPT Health vs. Claude for Healthcare vs. Google’s Fitbit AI Coach

Feature	ChatGPT Health (OpenAI)	Claude for Healthcare (Anthropic)	Fitbit AI Coach (Google)
Availability	All U.S. users 18+, all plans	Rolled out January 2026	Fitbit device owners
Data integrations	Apple Health, MyFitnessPal, Function, Epic, Oracle Health, One Medical	Enterprise/health-system focused	Fitbit activity & sleep data
Where it works	Any ChatGPT conversation	Health-specific workflows	Fitbit app
Clinical intent	Explicitly not for diagnosis/treatment	Positioned for clinical support contexts	Wellness/fitness coaching
Benchmark cited	HealthBench	Not publicly specified in the same way	Not applicable

Note: Feature sets and availability change quickly in this space — always check each provider’s current documentation before relying on comparisons.

Should You Use ChatGPT Health? Pros and Cons

Pros:

• Convenient — no need to switch apps to ask a quick health or nutrition question
• Can factor in your actual activity, diet, or medical history if you choose to connect it
• Backed by a dedicated benchmark and physician input, according to OpenAI
• Free to use at a basic level across all plan tiers

Also read;

OpenAI Wants ChatGPT to Become Your Next Super App – Here’s What That Could Mean

AI Skills Race Is Coming for the Auto Industry as Car Companies Rethink Jobs

Anthropic Pushes Claude Deeper Into Legal AI as Competition Heats Up

For Tech Blog

Cons:

• Not designed or approved for diagnosis or treatment
• Independent research suggests AI health advice can be inconsistent or wrong
• Connecting medical records means sharing sensitive data with a third-party AI system
• A live lawsuit alleges a serious real-world harm tied to health-related chatbot advice

Common Mistakes to Avoid

• Treating ChatGPT Health as a replacement for a doctor visit — it’s explicitly not designed for that.
• Connecting every available data source without reviewing what’s shared — be selective.
• Ignoring red-flag symptoms because a chatbot response sounded reassuring — always err toward professional care for anything serious or urgent.
• Assuming benchmark improvements equal clinical reliability — HealthBench measures model performance on structured evaluations, not real-world outcomes.

Best Practices for Using AI Health Tools Safely

• Use it for general education, not diagnosis — think “help me understand this term” rather than “tell me what’s wrong with me.”
• Always verify anything important with a licensed healthcare provider.
• Review your connected data sources periodically and disconnect anything you no longer want linked.
• If a response ever tells you not to see a doctor for a concerning symptom, treat that as a reason to seek care anyway, not a reason to skip it.

13. Key Takeaways

• ChatGPT Health is now live for all logged-in U.S. users 18+, across every plan, on web and iOS.
• It can now surface health context in any conversation, not just a dedicated hub — because most usage was already happening that way.
• Weekly health-related queries have grown from 230 million to 300 million.
• OpenAI says it doesn’t train models on this health data and cites HealthBench improvements with its newest small model.
• The rollout follows a lawsuit alleging harmful health advice, and independent research continues to flag reliability concerns with AI medical guidance.
• OpenAI’s own terms state the product isn’t intended for diagnosis or treatment — a caveat worth taking seriously.

Source

14. FAQs