ChatGPT Improvements for Clinicians

OpenAI has launched ChatGPT for Clinicians, offering it for free to verified doctors, NPs, PAs, and pharmacists in the US. It's designed to assist with clinical tasks such as documentation, medical research, and patient guidance, and takes a separate approach from ChatGPT for Healthcare, which is for organization-wide deployment.

The product includes skills for repetitive tasks, clinical search based on real-time evidence, deep research covering medical journals broadly, and CME-linked features recognized for real clinical questions. HIPAA support is also available via a BAA when needed, conversations are not used for training, and security measures such as MFA are applied.

To improve safety and accuracy, OpenAI has been continuously evaluating model responses in the medical field, with a panel of physician advisors having reviewed 700,000+ responses so far. Before launch, 6,924 conversations were tested, and physicians rated 99.6% of them as safe and accurate.

The newly released HealthBench Professional is a public benchmark covering 3 use cases: care consult, writing and documentation, and medical research. OpenAI stated that in this evaluation, GPT-5.4 performed better than base GPT-5.4, other OpenAI and external models, and human physicians.

The free version is currently available only in the US, with plans to expand to other countries and groups going forward. OpenAI also released a Health Blueprint, stating that the medical community, regulators, and health systems need to work together to advance responsible AI adoption.