HeyGen's Newest Avatar IV Model Creates AI Doppelganger in Seconds with One Photo + One Word

HeyGen released the Avatar IV model, which can generate realistic AI digital human videos with only a photo and voice/text input, supporting multi-angle input and singing function; the new model adopts an audio-driven expression engine based on diffusion model, which can automatically generate corresponding expressions and movements according to voice rhythm and emotions, with details including throat knots, chest undulation, etc.; paid users can generate 30-second videos, and free users can get 10-second video credits three times a month. Free users can generate 30-second videos, and free users have 10-second video credits three times a month. The application scenarios cover animation production, virtual pets, game dubbing and other fields.