OpenAI is rolling out an upgrade aimed at making ChatGPT a more capable source of health and wellness information, powered by a new model the company calls GPT-5.5 Instant.
According to OpenAI's blog, the update brings stronger reasoning, better handling of context, clearer communication, and evaluations informed by physicians. The company says the improvements show up in head-to-head testing: according to The Decoder, OpenAI's own comparative tests found the model now outscores answers written by doctors on accuracy, clarity, and completeness. The Decoder also reports that, per OpenAI, the error rate for health-related statements has dropped by 71 percent.
The stakes are large because the audience already is. Business Insider reports that 230 million people ask ChatGPT for health advice every week — and that OpenAI wants that number to grow. The new responses are not limited to paying customers, either: Search Engine Journal reports that OpenAI is bringing the improved health answers to the free version of ChatGPT.
A few caveats are worth keeping in mind. The accuracy claims come from OpenAI's internal testing rather than independent review, and "outscoring doctor-written answers" in a benchmark is not the same as replacing medical care. Still, the direction is clear: OpenAI is positioning ChatGPT as a go-to tool for health questions rather than treating the topic as something to steer users away from.
Why it matters: with hundreds of millions of people already turning to chatbots for health guidance each week, even small gains — or errors — in how an AI answers medical questions ripple out to a very large and vulnerable audience.