NewsBite analysis · AI, HealthTech
The Decoder
ChatGPT will give you worse health advice if you don't pay
OpenAI has launched its 'Health in ChatGPT' feature, providing varying levels of health advice depending on subscription status. Users on free plans receive less accurate information as the upgraded model, GPT-5.6 Sol, is only available to paying customers. Ethical considerations around AI advice versus human medical expertise underscore both the system's benefits and its shortcomings, particularly in less-than-ideal testing environments.
MonitorConfidence: MediumShort term horizonReputational risk: High
Published 23 Jul 2026, 20:10 UTC2 min read
Visit original article ↗Lead image — The Decoder
Summary
TL;DR
- Health advice quality varies between free and paying users of ChatGPT.
- GPT-5.5 Instant is used for free users, while GPT-5.6 Sol is for subscribers.
- OpenAI claims both models outperform average doctors in benchmark tests.
- Doctors may provide less accurate answers under pressure compared to AIs.
- The Health feature is not yet available in Europe due to regulatory concerns.
Interesting facts
- More than 300 million health-related inquiries are made weekly on ChatGPT.
- The Health feature was developed with input from over 260 physicians.
- Studies show AI performance in health can sometimes outperform human professionals.