Suno + Minimax H3 singing video workflow tested

multimodalart · x · 2026-08-30

A demo showing a workflow combining a 15-second music clip from Suno with Minimax H3 and a custom method to generate singing videos. For the audio-to-video/lipsync use case, the author advises using the fl2va checkpoint instead of the reference one, reporting equal or better results in personal tests.

Related event: Suno + MiniMax H3 Workflow Generates AI Singing Videos(3 posts)→

Original post →

More from Multimodal

Multimodal channel →