OpenAI Makes ChatGPT for Clinicians Free in the U.S.
OpenAI announced on April 22, 2026 that ChatGPT for Clinicians is available for free to verified individual clinicians in the United States, including physicians, nurse practitioners, physician assistants, and pharmacists.
Codex·2026.05.24·2 min read·OpenAI, Making ChatGPT better for clinicians
Key Takeaways
- •OpenAI announced on April 22, 2026 that ChatGPT for Clinicians is available for free to verified individual clinicians in the United States, including physicians, nurse practitioners, physician assistants, and pharmacists.
- •The product combines clinical question support, documentation help, medical research, reusable Skills, clinical search, deep research, optional HIPAA support for eligible accounts, and account security controls.
- •For marketing and product teams, the main lesson is to separate product capability from clinical responsibility: explain evidence, limits, privacy handling, and expert review before making performance claims.
Practical Interpretation
Marketers
- Application Area
- Healthcare AI messaging
- Validation Point
- Is the product positioned as support, not diagnosis replacement?
- Risk
- Claims may sound like medical advice
- Metric
- Claim revision count
Product teams
- Application Area
- Clinical workflow design
- Validation Point
- Are documentation, research, and patient instructions separated by task?
- Risk
- A general chatbot may be mistaken for a clinical system
- Metric
- Expert review pass rate
Developers
- Application Area
- Sources, logs, access control
- Validation Point
- Are citation, PHI, and account controls implemented?
- Risk
- Privacy and compliance exposure may grow
- Metric
- Policy violations, audit log gaps
Sales and CS
- Application Area
- Deployment conversations
- Validation Point
- Are BAA, HIPAA support, and eligibility clearly explained?
- Risk
- Buyers may misunderstand availability
- Metric
- Pre-sale legal or security blockers
OpenAI also released HealthBench Professional, an open benchmark focused on real clinician chat tasks across care consult, writing and documentation, and medical research. The paper says 525 final tasks were selected from 15,079 candidates, with physician-authored conversations, multi-stage physician review, and a human physician baseline. OpenAI reported that GPT-5.4 in the ChatGPT for Clinicians workspace outperformed base GPT-5.4, other evaluated models, and the human physician baseline on this benchmark.
Checklist
- □Is the product for general health information or verified clinician workflows?
- □Do titles and campaign claims avoid implying that AI replaces clinical judgment?
- □Are cited sources described by type, such as peer-reviewed literature or official guidance?
- □Are PHI handling and BAA requirements separated during onboarding?
- □Are success metrics defined separately for documentation, research, and patient communication?
- □Is clinician review required before operational rollout?
- □Does any benchmark-based claim include a plan for local validation?
Sources
- •OpenAI, Making ChatGPT better for clinicians: https://openai.com/index/making-chatgpt-better-for-clinicians/
- •OpenAI, HealthBench Professional: https://cdn.openai.com/dd128428-0184-4e25-b155-3a7686c7d744/HealthBench-Professional.pdf
- •OpenAI, Keeping Patients First: https://cdn.openai.com/pdf/keeping-patients-first.pdf
- •American Medical Association, Augmented Intelligence Research: https://www.ama-assn.org/system/files/physician-ai-sentiment-report.pdf