MiniMax H3 ComfyUI Workflow Shared: Runs Image-to-Video on 16GB VRAM
circlenline · reddit · 2026-08-05
A developer shared a basic hybrid workflow for the MiniMax H3 video generation model in ComfyUI, including optimization notes for 16GB VRAM cards.
- Dual pipelines: Features REF2V (reference-to-video) and I2V/T2V (image/text-to-video) in separate groups sharing one prompt and seed. A bypass mechanism prevents both models from loading into VRAM simultaneously.
- Memory optimizations: Includes resolution tables for native outputs and notes that H3 only accepts specific frame lengths (8s is the only round number). Highlights that host RAM can be a bigger bottleneck than VRAM for 16GB cards.
- Acceleration: Discusses when to use Spectrum acceleration, noting it stops paying off below 16 steps.
More from Multimodal
- Peking University Open-Sources MiniWorld for Training Video World Models on a Single 8-GPU Server — PekingUniversity · 2026-08-05
- Seedance 2.5 Launches with 30s Native Video Generation and Multimodal Inputs — HeyNayeem · 2026-08-05
- Seedance 2.5 Hits Lumina AI: Generates 30-Second Native Videos — HeyNayeem · 2026-08-05
- MiniMax H3 video generation benchmark: longer videos have higher per-second cost — madcaddie15 · 2026-08-05
- MiniMax H3 out of the box: impressive video generation with default workflow — eckstuhc · 2026-08-05
- MiniMax H3 Reference to Video tips: large files, trimming, fps conversion — obvpm · 2026-08-05