OpenAI boosts free ChatGPT health answers, claims frontier quality

Image credit: Search Engine Journal

OpenAI significantly enhanced the accuracy of health-related responses in its free ChatGPT model, GPT-5.5 Instant, claiming performance now rivals its advanced Thinking models.

The company reported that GPT-5.5 Instant, the default model for free ChatGPT users, now performs comparably to its frontier Thinking models on health questions, based on internal evaluations.

OpenAI stated it observed a 71 percent reduction in factuality problems on live traffic health responses flagged for at least one potential issue over a two-month period.

A panel of physicians rated GPT-5.5 Instant’s responses higher than physician-written ones on accuracy, communication, and completeness across 3,500 reviewed responses, according to the company.

The evaluations utilized OpenAI‘s HealthBench benchmark, which was developed with a network of more than 260 physicians across 60 countries. Doctors reviewed over 700,000 example responses as part of this process.

Health and wellness inquiries are among the most frequent reasons people use ChatGPT, with more than 230 million users posing such questions weekly, OpenAI reported.

The company confirmed it would not run advertisements in ChatGPT conversations concerning health, mental health, or politics, categorizing health as a protected topic.

These accuracy claims are based on OpenAI’s in-house tests and have not undergone independent verification by a third party.

The company’s assertions have significant implications for publishers and the broader scrutiny of AI-generated content, particularly in sensitive areas like health.

Competitors like Google also continue to develop their AI capabilities, with the broader industry facing increasing pressure to ensure the reliability of AI outputs.

Source: Search Engine Journal