Kandinsky Open-Sources Speech-to-Video Talking-Head Model S2V 1.0

Kandinsky Lab has open-sourced Kandinsky S2V 1.0, a speech-driven avatar video generation model available in Lite (2B) and Pro (21B) versions. It generates lip-synced talking-head videos from a single reference photo and an audio clip.

2026-10-16 ~ 2026-10-16 · 2 related posts

• Kandinsky Lab open-sources S2V 1.0: audio-driven avatar video from one still, MIT licensed — lmoroney · 2026-10-16

1 near-duplicate retellings: lmoroney