OpenAI has rolled out the GPT-5.5 Instant model as the default handler for all health-related queries across ChatGPT, including for users on the free tier, per the company’s official June 2026 health intelligence update. The update delivers a 71% reduction in flagged health response factuality issues, measured across the platform’s full health-related message volume. OpenAI’s official health intelligence announcement
How did OpenAI develop GPT-5.5 Instant’s health intelligence upgrades?
OpenAI developed the GPT-5.5 Instant health intelligence upgrades in partnership with a global network of 260+ licensed medical professionals. This group spans 60 countries, covers 49 languages, and represents 26 distinct medical specialties, per the company’s official announcement. OpenAI’s official health intelligence announcement
A physician-led evaluation tested the model across 3,500 representative health query scenarios. Researchers specifically compared GPT-5.5 Instant outputs to both prior ChatGPT model versions and human-authored physician responses to measure performance gaps. OpenAI’s official health intelligence announcement
The test found GPT-5.5 Instant missed fewer urgent care red flags, and more consistently requested additional user context when available information was insufficient to provide safe guidance. OpenAI’s official health intelligence announcement
This health intelligence work builds on OpenAI’s broader clinical AI research portfolio. For example, its validated rare childhood disease diagnosis tool supports clinicians in identifying conditions with limited existing diagnostic pathways. OpenAI rare childhood disease diagnosis research
How was the 71% factuality improvement validated?
OpenAI measured real-world model performance using production ChatGPT health query data. The platform processes billions of health queries per week, per the company’s official announcement. OpenAI’s official health intelligence announcement
Researchers tracked factuality issue rates across this full query volume, then specifically compared pre-update and post-update response data to quantify performance change. This analysis confirmed the 71% drop in flagged health response factuality issues for GPT-5.5 Instant, measured against the platform’s full health-related message volume. OpenAI’s official health intelligence announcement
Is the improved health model available to free ChatGPT users?
The GPT-5.5 Instant update is rolling out to all ChatGPT users, including those on the free tier. This expands access to the model’s improved health query handling across the platform’s full user base, rather than limiting the upgrade to paid subscription tiers only. OpenAI’s official health intelligence announcement
Bottom line: All ChatGPT users, including free tier account holders, now have access to the GPT-5.5 Instant model as the default handler for health queries, with a verified 71% reduction in flagged health response factuality issues measured across the platform’s full health query volume.
