ModelsSecurityBusiness

OpenAI upgrades health answers with GPT‑5.5 Instant — beats doctor‑written responses in tests

The key point: OpenAI rolls out GPT‑5.5 Instant to improve ChatGPT’s healthcare capabilities, matching expensive Thinking models on HealthBench tests and outperforming doctor‑written answers in evaluations.

In detail

  • GPT‑5.5 Instant is available to all free ChatGPT users, with usage limits
  • Matches performance of pricier Thinking models on HealthBench and HealthBench Professional
  • Compared with doctors, responses scored higher on accuracy, clarity, completeness; incorrect health statements dropped 71% over two months
  • Improvements driven by a network of 260+ doctors from 60 countries who reviewed over 700,000 responses; OpenAI cites 230M weekly users for health queries

Why it matters

Higher‑quality, lower‑cost health responses in a widely available model change how businesses provide basic medical information and triage — offering efficiency gains but raising liability and safety considerations.

For you Actionable takeaway: Pilot GPT‑5.5 Instant for non‑clinical health guidance, define clear escalation paths to clinicians, and consult legal/compliance before public use.

← All news

Summaries are generated automatically and link to the original source.