Is ChatGPT for Clinicians free, and who can use it?
Yes, it is free for verified clinicians in the United States, and it is a separate workspace, not a new plan. OpenAI's help center says: "ChatGPT for Clinicians is free for verified clinicians in the United States." Its release notes date the launch to 22 April 2026. Here is who gets in, as of 15 November 2026:

| Role | Eligible | Source line |
|---|---|---|
| Physician (MD or DO) | Yes | Listed by OpenAI |
| Nurse practitioner (NP) | Yes | Listed by OpenAI |
| Physician assistant (PA) | Yes | Listed by OpenAI |
| Pharmacist | Yes | Listed by OpenAI |
| Registered nurse, dentist, therapist, student | Not listed | The help article names four roles and no others |
| Clinician outside the US | Not at launch | OpenAI plans "to expand access to additional countries over time" |
Free comes with a caveat in OpenAI's own wording. The FAQ says: "At launch, ChatGPT for Clinicians is free for verified clinicians in the United States." OpenAI hasn't published a date when that changes. Here is what OpenAI reports about its pre-launch testing, on a log scale:
Those numbers come from OpenAI's launch post, not from an outside audit. It reports that across 6,924 test conversations, physicians rated 99.6% of responses as safe and accurate. That is OpenAI's own test, and it doesn't change what the next section says about who is responsible for the answer.
How do you sign up for ChatGPT for Clinicians?
You sign up in about five steps with a ChatGPT account and your NPI. OpenAI's help center says you need three things first: a ChatGPT account, a valid NPI, and a license the third-party verification provider can check. Then, per the same article:
- Go to the ChatGPT for Clinicians page at chatgpt.com/cfc-plan.
- Sign in with your existing ChatGPT account, or create one.
- Complete clinician verification through the third-party provider, using your NPI.
- Confirm that you are a licensed clinician and agree to the services agreement.
- Select Get started with ChatGPT to create your ChatGPT for Clinicians workspace.
Your old workspace stays. OpenAI says it "remains available", and you switch between the two from the account switcher. That matters because the clinician workspace has its own settings, its own BAA and its own credits, and a clinician who pastes into the wrong workspace has skipped the agreement without noticing. Check the switcher first.
What does ChatGPT for Clinicians include?
It includes trusted clinical search with citations, deep research, pre-built skills, documentation drafting and CME credit, plus the usual ChatGPT tools. OpenAI's launch post says it builds on models like GPT-5.4. The table is compiled from the help center and launch post on 15 November 2026:
| Feature | What OpenAI says | Limit to know |
|---|---|---|
| Trusted clinical search | Answers drawn from peer-reviewed studies, public health guidance and clinical guidelines, with titles, journals, authors and dates | ChatGPT picks the sources; you check them |
| Deep research | Cited reports across medical literature | OpenAI says to review the sources before relying on it |
| Skills and starter prompts | Reusable workflows for referral letters, prior authorization and patient instructions | Drafts only, for your review |
| CME credit | Eligible clinical questions can count toward credit | Only "eligible" questions count |
| Custom GPTs, projects, connected apps, memory, canvas | Supported | Third-party GPTs and apps have their own terms |
| Healthcare Public Data plugin | Nine read-only apps, including PubMed, ClinicalTrials.gov, DailyMed and openFDA | Added 1 September 2026; no PHI in these queries |
| Image generation | Not supported | OpenAI lists it as unsupported |
| Epic plugin | Not available | Only on ChatGPT for Healthcare |
OpenAI's help article names no message cap for the clinician workspace. It doesn't say there is none either, so treat the limits as unpublished as of November 2026. If you hit a wall, the Work and Codex credit notes in the FAQ say that purchases made in a separate personal workspace won't add credits here.
Is ChatGPT for Clinicians HIPAA compliant?
It supports HIPAA use only if a Business Associate Agreement is in place, and the clinician has to sign it. OpenAI's launch post says "HIPAA support is available through a Business Associate Agreement (BAA) for eligible accounts." The HIPAA page lists ChatGPT for Clinicians among six HIPAA eligible products. Free, Go and Plus aren't on that list.
Here is the rule in OpenAI's words: "Do not share PHI in ChatGPT for Clinicians unless a BAA is in place and you are authorized to sign a BAA for your account." Eligible individuals review and sign it in ChatGPT under Settings, then Agreements. If you work for a hospital, the clinician workspace isn't the place to decide that. Ask whether your employer already has a BAA, because OpenAI points anyone who needs a BAA covering several users to ChatGPT for Healthcare.
A signed BAA doesn't cover every feature. OpenAI's HIPAA page, read 15 November 2026, says features not listed as covered aren't automatically covered. Its not-covered list includes internet access in Canvas and code blocks, browser use for Work in the cloud, network access for Work in the cloud, event-triggered scheduled tasks, workspace agents with agent-owned connections, and improved memory (context from past chats). Memory is a clinician-workspace feature, so check that setting before PHI goes in. Third-party GPTs and connected apps follow their own terms, and OpenAI says not to send them PHI without the right agreements.
Many clinical tasks don't need PHI at all, which OpenAI says in its launch post. OpenAI also says content shared with the workspace isn't used to train its models. A BAA is a contract between you and OpenAI that sets how patient data is handled, and it isn't a license to paste a full chart. Keep it small.
What must you not use ChatGPT for Clinicians for?
Don't use it to replace your own judgment, and don't use it for anything that sends PHI where no BAA reaches. This page isn't medical advice. It reports what OpenAI's pages say, as of 15 November 2026, and OpenAI's own limits are the ones that count:
- Care decisions. The help article says it "is intended to support, not replace, professional judgment", and that "Clinicians remain responsible for all care decisions and should independently verify information as appropriate."
- Patient data without a BAA. No PHI unless you signed the agreement for your account.
- Third-party GPTs and apps. No PHI unless you hold the right permissions and agreements.
- Public data searches. OpenAI says not to put patient names, record numbers or other PHI into Healthcare Public Data queries.
- Clinical-trial and coverage answers. Confirm trial openings with the study team. OpenAI says Medicare coverage information "does not determine an individual patient's benefits."
- Provider checks. The NPI Registry app doesn't prove licensure or quality, per OpenAI.
The workspace is also for clinicians, not for patients. A patient reading this page should talk to their own care team. Whizi, which publishes this page, can't answer a medical question for you either.
How is it different from ChatGPT for Healthcare and a normal plan?
ChatGPT for Clinicians is one person on a free self-serve workspace, while ChatGPT for Healthcare is an organization on an enterprise contract. OpenAI's help center says Healthcare is for deploying ChatGPT "across clinicians, administrators, and researchers with centralized administration, company-specific data connections, and enterprise security and compliance controls." The comparison is built from OpenAI's pages, read 15 November 2026:
| ChatGPT for Clinicians | ChatGPT for Healthcare | ChatGPT Free, Go, Plus | |
|---|---|---|---|
| Who buys it | An individual clinician | A health system or organization | An individual |
| Price | Free "at launch" | Based on ChatGPT Enterprise; contact sales | Public prices |
| BAA | Individual signs in Settings | Organization-wide, through OpenAI | Not on the HIPAA eligible list |
| Admin controls | None named | Role-based access, SSO, SCIM, audit logs | None |
| Epic plugin | Not available | Yes, set up by an admin | No |
| Clinical search | Yes | Yes, enabled by an admin | No |
| Who qualifies | Verified MD/DO, NP, PA, pharmacist in the US | Organizations | Anyone |
Pick Clinicians if you are one eligible clinician who wants cited answers and drafting. Pick Healthcare if a team needs one agreement, admin controls or Epic. Individual clinicians can also hold a Clinicians workspace when their employer runs Healthcare, per the FAQ. If you're a registered nurse, AI for nurses covers what to do when the list doesn't include you.
Where does Whizi fit for a clinician?
Whizi isn't a clinical tool and isn't HIPAA compliant, and you shouldn't paste patient information into it. It has no BAA. If you're eligible, ChatGPT for Clinicians is the better home for anything touching care, because it is free and it offers a BAA. Whizi earns a place for the work with no patient in it: study notes, patient handouts written from scratch, emails, and reading a paper outside the clinical workspace. ChatGPT for Clinicians is ChatGPT only, so Claude and Gemini sit elsewhere. Whizi Pro puts Claude, GPT and Gemini in one app, and its first week is $1.99. The free plan gives 10 messages a day on Whizi AI, Auto and Claude Haiku 5.5, no card. What never to paste into AI lists what stays out of every app.
- Check that your role is one of the four OpenAI lists: MD/DO, NP, PA or pharmacist
- Have your NPI and a license the verifier can check before you start at chatgpt.com/cfc-plan
- Switch to the clinician workspace in the account switcher before you chat
- Read the BAA under Settings, then Agreements, before any PHI goes in
- Keep PHI out of third-party GPTs, connected apps, Canvas internet access and public data searches
- Verify every cited source yourself, because you stay responsible for the care decision
Frequently asked questions
Is ChatGPT for Clinicians HIPAA compliant?
It is HIPAA eligible, which isn't the same as automatically compliant. OpenAI lists it among six HIPAA eligible products, and an eligible individual can sign a BAA under Settings, then Agreements. OpenAI says not to share PHI unless that BAA is in place, and some features sit outside it. What never to paste into AI covers the line.
Can nurses use ChatGPT for Clinicians?
Not if they are registered nurses, as of 15 November 2026. OpenAI's help article lists physicians, nurse practitioners, physician assistants and pharmacists, so NPs qualify and RNs aren't named. AI for nurses explains how to use AI without patient data.
Does ChatGPT for Clinicians work outside the US?
No, not at launch. OpenAI says access is for verified clinicians in the United States and that it will begin piloting outside the US with the Better Evidence Network. No country or date is named. The clinician workspace needs a US NPI to sign up.
How much does ChatGPT for Clinicians cost?
Nothing, for eligible verified US clinicians, "at launch" in OpenAI's words. OpenAI hasn't published when or if that changes. ChatGPT for Healthcare is priced from ChatGPT Enterprise and sold by contact with OpenAI's sales team. Best AI for medical questions compares other tools.
Does OpenAI train on what you type into ChatGPT for Clinicians?
No, according to OpenAI. Its help center says content shared with ChatGPT for Clinicians is not used to train OpenAI's models. That is separate from the BAA, which you still have to sign before any PHI goes in, and separate from third-party GPTs and apps, which have their own terms.
Does ChatGPT for Clinicians have usage limits?
OpenAI's help article publishes no message cap for the clinician workspace, as of 15 November 2026. It does say Work and Codex can show an out-of-credits message, and that credits bought in a separate personal or Business workspace don't carry over.