OpenAI reserves stronger health advice for ChatGPT premium subscribers
OpenAI is tiering health advice quality by subscription status, reserving its more capable GPT-5.6 Sol model for paying subscribers while free users receive answers from the weaker GPT-5.5 Instant. The rollout integrates Apple Health, medical records, and wellness apps into ChatGPT, capitalizing on the 300 million weekly health queries already flowing through the platform. This move crystallizes a broader industry pattern: monetizing capability gaps rather than access alone, and raises acute questions about equity in AI-mediated healthcare guidance where financial status directly determines answer reliability.
Modelwire context
Analyst takeThe detail worth sitting with is that the gap isn't cosmetic. Routing health queries to a materially weaker model for free users means the reliability floor for medical guidance is now explicitly set by willingness to pay, not by platform-wide safety standards. OpenAI is, in effect, treating health accuracy as a premium feature rather than a baseline obligation.
This is largely disconnected from recent activity in our archive, as we have no prior coverage to anchor it to. But it belongs squarely in the broader conversation about how AI companies are transitioning from access-based monetization (free vs. paid) to capability-based monetization (worse vs. better answers for the same question). That shift has different stakes in health contexts than in, say, coding assistance, because the downstream cost of a degraded answer is not a broken script but a missed symptom or a delayed care decision. The Apple Health integration amplifies this by pulling in personal medical records, raising the consequence of any model-quality shortcut.
Watch whether any US state attorney general or the FTC opens an inquiry into whether tiered health advice quality constitutes an unfair or deceptive practice under consumer protection statutes. If a formal investigation is announced within six months, it signals regulators are willing to treat model-tier decisions as a health equity issue, not just a product choice.
This analysis is generated by Modelwire’s editorial layer from our archive and the summary above. It is not a substitute for the original reporting. How we write it.
MentionsOpenAI · ChatGPT · GPT-5.6 Sol · GPT-5.5 Instant · Apple Health
Modelwire Editorial
This synthesis and analysis was prepared by the Modelwire editorial team. We use advanced language models to read, ground, and connect the day’s most significant AI developments, providing original strategic context that helps practitioners and leaders stay ahead of the frontier.
Modelwire summarizes, we don’t republish. The Decoder originally reported this story as “ChatGPT will give you worse health advice if you don't pay”. The full content lives on the-decoder.com. If you’re a publisher and want a different summarization policy for your work, see our takedown page.