OpenAI releases computer-use agents for healthcare with 99% clinical safety rating

·2026.11.16 15:00

Key point

GPT-6 Astra scored 94.3% on HealthAdminBench while GPT-6 Luna achieved 91.8% at 98% lower cost per task.

Details

  • OpenAI launched computer-use agents for healthcare teams, available via the OpenAI API, ChatGPT for Healthcare, and ChatGPT for Clinicians, supporting HIPAA-compliant workflows with audit logs.
  • In clinical evaluations, 99% of 325 reviewed tasks were rated safe, and physicians rated the work as "very helpful" or "extremely helpful" in 93% of 646 assessments.
  • On HealthAdminBench, GPT-6 Astra earned 94.3% of rubric points, while GPT-6 Luna earned 91.8% at a cost of 8.5 cents per task, which is 98% less than Astra.
  • Partners including CVS Health, Abridge, Cigna Healthcare, and Ambience are deploying these agents to automate administrative tasks like prior authorizations, referrals, and discharge planning.

Deployment and Safety

  • CVS Health uses agents to clear document backlogs and process tens of thousands of consults daily, with data remaining within their enterprise-controlled infrastructure.
  • Cigna Healthcare reduced the time case managers spend navigating more than 10 separate systems to gather member context.
  • Ambience built controls to ensure every agent action is auditable, allowing health systems to pause agents and review step-by-step logs.
  • Organizations must use eligible services covered by a Business Associate Agreement with OpenAI to handle protected health information.

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.