
MAGI-2 Preview
Magi-2 Preview is a 114B-parameter unified audio-video generation model that activates just 6B parameters per token. Built on MagiMoE and co-designed across architecture, systems, and data, it explores an efficient path to scaling video generation.

MAGI-2 Preview
Character Performance
Precisely generates dialogue, singing, and emotional shifts; natural coordination of lip-sync, facial expressions, eye movements, and body language creates a more compelling performance.
Creative Partners








































