Blinded physicians find fewer flaws in a model's answers than in doctors'

Claim

GPT-5.6 set a new bar on professional health benchmarks, with blinded physicians finding fewer flaws in its answers than in specialty-matched doctors' own, while its smallest variant beat GPT-5.5's best reasoning at 25 times lower cost.

Reported metric: 25x cheaper

Reporting date: 2026-07-12 (not necessarily the event date)

Sources

Attribution and reporting linked by this record; links do not imply independent verification.

• Welcome to July 12, 2026 (theinnermostloop.substack.com)