OpenAI aims to reduce the time required for medical research to reach clinical practice from the current 5–17 years to hours or minutes. The company utilizes its general-purpose AI models and the ChatGPT for Clinicians product to convert complex medical literature into actionable guidance for physicians.
Launched in April 2026, ChatGPT for Clinicians provides verified US doctors with access to advanced models and HIPAA-compliant options. During pre-launch testing, the tool underwent approximately 7,000 conversations, achieving a 99.6% safety and accuracy rating.
The company developed HealthBench to evaluate AI performance in clinical settings. This benchmark incorporates 48,562 criteria written by 262 physicians from 60 countries and 26 specialties to assess 5,000 realistic medical conversations.
A study conducted with Penda Health in July 2025 across 39,849 patient visits demonstrated a 16% reduction in diagnostic errors and a 13% decrease in treatment errors. Additionally, OpenAI reports that over 300 million people use ChatGPT for health-related queries weekly.
