Free ChatGPT gives worse health advice than paid tier
openai launched 'health in chatgpt' powered by the gpt-5.6 sol model for paid subscribers, while free users get the inferior gpt-5.5 instant, creating a two-tier system where wealth dictates access to superior ai medical insights.
openai is rolling out 'health in chatgpt' to us users, allowing data integration from apple health and medical records to analyze lab results and prepare for appointments. the feature employs a two-tier model system: paid subscribers receive the flagship gpt-5.6 sol, while free users get the lower-scoring gpt-5.5 instant, which lags in decision helpfulness and completeness. although the top model surpassed physician answers on internal benchmarks, external studies like radle 2.0 show ai models still hallucinate with high confidence, posing risks for unsupervised medical use. this rollout highlights growing tensions between ai utility in healthcare and the ethical implications of monetizing health-related accuracy.