Unverified50% confidenceBenchmarkExact time
In HealthBench, doctors first wrote the grading rubric, and OpenAI's GPT-4.1 evaluator exceeded the average doctor's level on 5 of 7 topics.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/14/2026
First Seen
Valid until: 1/12/2027
Sources
Related Entities
Related Claims
UnverifiedOpenAI's medical evaluation had specialist physicians write answers, which were then blindly assessed by other physicians and scored across three dimensions: accuracy, communication quality, and helpfulness for health decision-making.69% similarUnverifiedFrontier general-purpose large models scored higher than the physician control group in expert evaluations, particularly excelling in medical advice, treatment principles, and certain diagnostic tasks.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1026036API
curl https://kongchang.com/api/v1/knowledge/claims/1026036MCP
get_claim(id=1026036)