Transluce Evaluates 77 AI Models on Mental Health Crisis Response

Transluce released the broadest independent evaluation to date of how AI systems respond to users in mental health crises, covering 77 model variants from OpenAI, Anthropic, Google, Meta and others. Former OpenAI executive Neil Chowdhury called it the most thorough independent behavioral analysis in real multi-turn settings.

2026-09-01 ~ 2026-09-01 · 4 related posts

• Transluce evaluates 77 AI models on mental health crisis response — cogconfluence · 2026-09-01
• Transluce's mental health eval hailed as the most thorough independent model analysis — Miles_Brundage · 2026-09-01

2 near-duplicate retellings: cjmaddison · Miles_Brundage