HeyGen Launches Avatar V — 15-Second Video Creates Photorealistic AI Digital Twin with Micro-Expression Modeling
HeyGen·medium signal
HeyGen released Avatar V, claiming the most realistic AI avatar model yet. It creates a digital twin from a 15-second video clip, modeling dental structure, skin texture, facial geometry, and behavioral patterns like talking rhythm and micro-expressions. Unlike prior systems that condition on a single reference frame, Avatar V operates over a full video context window with selective attention for identity signals. Supports 175+ languages with voice cloning. Available on all HeyGen paid plans.