OpenAI Says Its New ChatGPT for Doctors Outperforms Humans in Clinical Tasks

OpenAI connected Wednesday unveiled a free, specialized mentation of ChatGPT for physicians, caregiver practitioners, doc assistants, and pharmacists, deepening the company's propulsion into a healthcare assemblage grappling with staffing shortages, administrative overload, and burnout.

The product, called ChatGPT for Clinicians, is designed to grip documentation, aesculapian probe and attraction consultations—time-consuming tasks that devour into the hours clinicians tin walk with patients. Access is presently constricted to verified practitioners successful the United States, with plans to grow internationally.

The announcement arrives arsenic AI adoption successful medicine is surging. According to a 2026 survey by the American Medical Association cited by OpenAI, 72% of physicians present usage AI successful objective practice, up from 48% conscionable a twelvemonth ago. The institution says clinician usage of its ain level has much than doubled implicit the past year, with millions relying connected ChatGPT weekly.

Among the tool's features are a objective hunt relation drafting connected millions of peer-reviewed sources, a heavy probe mode for aesculapian lit reviews, reusable workflow templates for tasks specified arsenic referral letters and anterior authorization requests, and the quality to gain continuing aesculapian acquisition credits portion researching objective questions successful the platform.

Conversations volition not beryllium utilized to bid OpenAI's models, and HIPAA compliance enactment is disposable done a Business Associate Agreement for eligible accounts.

Alongside the launch, OpenAI released HealthBench Professional, a caller benchmark designed to measure AI show connected realistic objective tasks crossed 3 categories: attraction consultations, documentation, and aesculapian research.

The institution reported that GPT-5.4, moving successful the ChatGPT for Clinicians workspace, scored 59.0 connected the benchmark—higher than quality physicians, who scored 43.7 adjacent with unlimited clip and net access, and higher than competing models from Anthropic, Google, and xAI.

Those results, however, travel with an important caveat: OpenAI built some the merchandise and the benchmark utilized to measure it.

To make the tool, the institution says it worked with hundreds of doc advisors and reviewed much than 700,000 exemplary responses. In pretesting, physicians rated 99.6 percent of responses arsenic harmless and close crossed astir 7,000 conversations.

OpenAI has been cautious to framework the instrumentality arsenic a enactment strategy alternatively than a replacement for objective judgment—a favoritism regulators and skeptics volition apt ticker intimately arsenic the merchandise rolls retired much broadly.

Healthcare represents a rapidly expanding marketplace for AI tools. Beyond OpenAI’s information showing that clinician usage of ChatGPT has much than doubled implicit the past year, McKinsey data indicates 50% of healthcare leaders study their organizations person implemented generative AI, up from 47% successful Q4 2024 and 25% successful Q4 2023. BCG research, meanwhile, shows 60% of consumers already usage AI for idiosyncratic health.