Minimax H3 in Practice: Ref2VA Workflow Generates 10s Clips in 2-4 Minutes
FewApple7788 · reddit · 2026-09-03
A hands-on workflow for Minimax H3 video generation: Ref2VA with 2-4 reference images per run, scale set to 0.7, outputting 1152×640. On a 5090 with 64GB RAM, each 10-second clip takes 2-4 minutes depending on image and audio refs (using saga). The author is happy with the results.
More from Multimodal
- Ant open-sources 5 Ming-3.0 speech models spanning TTS, audio understanding and ASR — Med1_Ai · 2026-10-21
- What keeps AI videos watchable: hooks, consistency and a payoff — aftahi_ai · 2026-10-21
- MORPHJUTSU skill in Flova transfers any video's motion to your character — aftahi_ai · 2026-10-21
- NVIDIA open-sources MotionBricks: 350K motion skills at 15,000 FPS, 2ms latency — aftahi_ai · 2026-10-21
- Runway joins Claude Marketplace, letting enterprises spend Anthropic budget on video generation — runwayml · 2026-10-21
- Kokon 0.10 turns Qwen-Image 2.1's text encoder into a prompt writer, no extra model — Ok-Director4174 · 2026-10-21