MiniMax H3 Open-Weight Video Model Now in ComfyUI, Supports 15s 768p Clips
pmttyji · reddit · 2026-08-10
MiniMax H3 is an open-weight multimodal video generation model handling text, image, video, and audio. In ComfyUI, it enables text-to-video, image-to-video, first/last-frame, and reference-driven creation. H3 jointly generates visuals and synchronized stereo audio (dialogue, SFX, ambience, music) rather than adding audio later. Open weights support clips up to 15 seconds at 768p; hosted version supports up to 2K. The post discusses unified architecture, high-compression video representation for efficiency, and local setup tips.
Related event: MiniMax Launches H3 Multimodal Video Model with ComfyUI Support(2 posts)→
More from Multimodal
- ForcingBooster halves computation of autoregressive video diffusion with patch-wise denoising — VAST-AI · 2026-10-16
- Tokyo's Gemini AI Photo Booth turns visitors into plush toys and dioramas — bradneuberg · 2026-10-16
- Orchestrating a 100-image portrait of humanity for aliens with ChatGPT, Argon and p5.brush — dx8xb · 2026-10-16
- Indie Creator Builds an AI-Native Animated Show Solo — ComprehensiveLime654 · 2026-10-16
- Ideogram teases 4.5 arrival and open weights: 'Not yet, but soon' — East_Shoe8814 · 2026-10-16
- Opus 5.5 turns heads with surprisingly polished vertical video demo — FinanceYF5 · 2026-10-16