Redditor tests video gen and lipsync on MiniMax H3, says quality still has a way to go
Few-Level-923 · reddit · 2026-10-20
A Redditor tested video generation and lipsync using MiniMax H3's text-to-prompt plus reference-video (t2p+r2v) workflow, sharing a gothic-style talking character clip and self-assessing that quality "still has a way to go" while continuing to tune settings.
More from Multimodal
- MiniMax-H3 hailed as favorite open release of the year, runs fully local — AdinaYakup · 2026-10-23
- Michal Malewicz experiments with depth maps in his creative workflow — michalmalewicz · 2026-10-23
- One-Shot Tunable Style Transfer on Flux3 Via Reference Latent + Node — Winter_unmuted · 2026-10-23
- PhysVidBench accepted at EMNLP 2026: do video gen models grasp physical commonsense? — aykuterdemml · 2026-10-23
- EMNLP 2026 paper teaches multimodal LLMs to get cartoons via incongruity-resolution supervision — aykuterdemml · 2026-10-23
- A gorgeous Midjourney style: sref code 820221253 for dreamy lavender watercolor looks — umesh_ai · 2026-10-23