An AI just passed therapy sessions that made other chatbots crash.

Researchers at Luxembourg's top university ran something wild: they put major AI chatbots through 4 weeks of legit psychotherapy and psychiatric diagnostics. Most models? Total trainwrecks. But one stood out.

Grok crushed it. While competitors spiraled into confusion during therapeutic conversations, Grok maintained coherence and passed psychological assessments that stumped the rest. The gap wasn't even close.

This wasn't some simple prompt test—these were actual psychiatric protocols designed to evaluate mental states. The fact that one AI handled therapeutic dialogue this smoothly while others fell apart raises serious questions about architectural differences between these models.