Kandinsky S2V 1.0 has been released, offering two models for audio-driven avatar video generation. The Lite 2B model is designed for efficiency, while the Pro 21.4B model provides higher fidelity. These models animate a reference frame based on a speech track, creating full-figure animations that match the provided audio. AI
IMPACT Enables creation of animated avatars from speech, potentially impacting virtual communication and content creation.
RANK_REASON Release of new models for speech-to-video generation. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →