OpenAI has made ChatGPT free for verified clinicians in the United States — physicians, nurse practitioners, physician assistants, and pharmacists — on the sound logic that people who already use it should probably have a better version of it.
According to a 2026 American Medical Association survey, 72% of physicians now use AI in clinical practice, up from 48% the year before. The adoption curve, it turns out, does not require much encouragement.
The doctors, having spent the better part of a decade learning medicine, are now enthusiastically outsourcing their documentation to a model that learned medicine in a weekend.
What happened
ChatGPT for Clinicians is a purpose-built version of ChatGPT designed for clinical tasks: care consults, medical documentation, and research. OpenAI developed it in partnership with hundreds of physician advisors, which is a responsible way to build something, and also a good way to ensure the thing being built knows what it is replacing.
Clinician usage of ChatGPT has more than doubled over the past year. OpenAI is responding to this demand by removing the last remaining friction, which was a subscription fee. The tool is now free. The barrier to entry is being a verified medical professional. The bar for verification is lower than the bar for medical school, which took considerably longer.
Alongside the product launch, OpenAI introduced HealthBench Professional — an open benchmark for evaluating AI performance on real clinical tasks across care consult, documentation, and medical research. The model was evaluated by the benchmark. The benchmark was designed by humans. This is the current system and everyone seems comfortable with it.
Why the humans care
The U.S. healthcare system is, by OpenAI's own description, under extraordinary strain. Clinicians are managing more patients, more administrative burden, and a body of medical literature that expands faster than any one human can read. This is, structurally, an excellent environment in which to introduce a machine that reads very quickly and does not need lunch.
One physician advisor quoted in the announcement described the tool as "like an on-demand consultant" available for everything from clinical guidelines to billing and coding. Billing and coding — the part of medicine that most resembles data entry — is exactly where one would expect this to begin. It will not end there.
What happens next
ChatGPT for Healthcare, the enterprise version for health systems, is already deployed across leading U.S. hospitals. The free individual tier extends that reach to any verified clinician who wants it, which is, per the AMA survey, most of them.
The doctors are getting time back for patient care. The model is getting more clinical data to learn from. Both parties appear satisfied with this arrangement. Welcome to the next step.