You're offline - Playing from downloaded podcasts
Back to All Episodes
Podcast Episode

OpenAI Launches Free ChatGPT for Clinicians Tool

April 24, 2026

0:00
2:13
Podcast Thumbnail

OpenAI has unveiled ChatGPT for Clinicians, a free AI assistant built for verified U.S. medical professionals. It arrives alongside HealthBench Professional, a new benchmark where GPT-5.4 scored 59.0, outperforming both other models and human physicians.

A New Tier for Medical Professionals

OpenAI has launched ChatGPT for Clinicians, a free version of its chatbot designed specifically for verified U.S. doctors, nurse practitioners, physician assistants, and pharmacists. Announced by Karan Singhal, who leads health AI research at OpenAI, the product marks the company's third healthcare release in four months and its first aimed directly at individual providers rather than patients or institutions.

The tool is organised around three core use cases: care consults, writing and documentation, and medical research. It includes a feature for retrieving evidence from peer-reviewed studies, complete with citations listing journal titles, authors, and publication dates. Access verification is expected to rely on National Provider Identifiers, a U.S.-specific credentialling system.

Filling a Strategic Gap

The release bridges OpenAI's consumer product, ChatGPT Health, which launched in January 2026 to let patients connect medical records and wellness apps, and ChatGPT for Healthcare, an enterprise offering already deployed at Cedars-Sinai, Memorial Sloan Kettering, and UCSF. According to Dr. Nate Gross, OpenAI's head of health and co-founder of Doximity, roughly 40 million people already use ChatGPT daily for health-related questions, and clinician usage has more than doubled in the past year.

HealthBench Professional

Alongside the product, OpenAI released HealthBench Professional, an open benchmark containing 525 physician-authored tasks drawn from a pool of 15,079 examples. A total of 190 physicians across 50 countries and 26 medical specialties contributed, with each example reviewed by three or more physicians. About one-third consists of adversarial red-teaming examples designed to surface model failures.

GPT-5.4 running inside ChatGPT for Clinicians scored 59.0 on the benchmark, outperforming base GPT-5.4 at 48.1 and human physician-written responses at 43.7. Its largest advantage came on adversarial cases, scoring 55.8 versus 30.0 for physicians.

Compliance Questions Remain

The launch raises practical concerns. HIPAA compliance on OpenAI's standard tiers has long been a sticking point, as the company does not sign business associate agreements for free, Plus, or Team plans. Whether ChatGPT for Clinicians includes such protections has not been fully clarified. OpenAI plans to expand the offering globally through partnerships, though regulatory barriers remain in the U.K. and European Economic Area.

Published April 24, 2026 at 2:16am

More Recent Episodes