Over 260 Doctors Across 60 Countries Evaluated 600,000 ChatGPT Health Responses Over Two Years

OpenAI involved more than 260 practicing physicians worldwide to assess ChatGPT Health's outputs. These experts reviewed over 600,000 model responses across 30 medical fields to help refine the product before its 2026 release.

OpenAI has disclosed extensive physician involvement in developing its ChatGPT Health product, which was publicly announced on January 7, 2026. More than 260 practicing doctors representing dozens of specialties from 60 countries evaluated upwards of 600,000 responses generated by ChatGPT Health over a two-year period. This collaborative effort focused on 30 different medical areas.

ChatGPT Health operates as a distinct interface within the broader ChatGPT platform, designed to integrate personal medical records and fitness app data. However, OpenAI emphasizes that the product is not intended for diagnosis and does not replace professional medical advice. Conversations within ChatGPT Health are excluded from training data for the base language models.

The feedback from these healthcare professionals has directly influenced how the system formulates medical recommendations, determines when to advise consulting a human specialist, and manages potentially hazardous or confusing outputs. Alongside ChatGPT Health's development, OpenAI created HealthBench, a benchmarking tool for assessing medical response quality, which was also developed with input from the same physician contributors.

It is important to note the distinction made by medical teams between expert model evaluation and clinical validation; OpenAI has not published any outcomes relating to real patient health results following interactions with ChatGPT Health. As such, the product remains a reference tool with medical expert calibration rather than a medical advisor or diagnostic system. ChatGPT Health became available in July 2026 to select U.S. users aged 18 and over via web and iOS platforms.