Tavus' real-time video model Griffin fooled 48% of participants in human test

thione · x · 2026-10-06

Tavus introduced Griffin, a real-time video-to-video "Human Interaction Model" that transforms live video into a photoreal driven persona.

Related event: Tavus Unveils Griffin: Real-Time Video AI Fooled 48% Into Thinking It Was Human(2 posts)→

Original post →

More from Multimodal

Multimodal channel →