The claim record · all entities

MentalHealthBench

The permanent record for MentalHealthBench. Every claim we have put on the ledger, checked against independent sources, with the ruling and the date. Each new ruling updates the counts below.

0 BS · 0 contested · 1 true, but · 0 verified

0 BS · 0 contested · 1 true, but · 0 verified, of the 1 claim from MentalHealthBench we have checked. This counts rulings. It is not a score. How we rule →

Last ruling: 25 Sep 2026

Most recent on the recordTRUE, BUTSelf-MarkedOpenAI built a mental health test, had its own model grade it, and its own model came top. Answers written by licensed clinicians scored 38.5%. GPT-6 Astra scored 57.3%.25 Sep 2026 · 13 sources

The full record 1 claim, newest first

September 2026 1 ruling

25 SepTRUE, BUTSelf-MarkedOpenAI built a mental health test, had its own model grade it, and its own model came top. Answers written by licensed clinicians scored 38.5%. GPT-6 Astra scored 57.3%.capability · 13 sources

Checked by AI at the BS Killer desk · latest ruling 25 Sept 2026 · How we check →

SEARCH EVERY MENTALHEALTHBENCH CLAIM IN THE LEDGER →
← ALL ENTITY DOSSIERS