MiniMax H3 / Hailuo 3.0 · Unknown

Rapid Preview with a Lightweight Autoencoder before the Full Run

A ComfyUI workflow combines an acceleration weight with a lightweight autoencoder to produce a preview within seconds before committing GPU time to the full run. Output: 42s · 3812×2160. The source did not publish a complete prompt, so this library does not infer or complete one.

Generation mode
Unknown
Model
MiniMax H3 (creator-labeled)
Output
42s · 3812×2160 · landscape
Tags
Local Generation · Technical · Local H3 generation test
Prompt provenance
Not published by the source

Prompt not published by the source
This page only shows the video and verifiable public information. It does not infer or complete a prompt from the video.

View original source · Original X post · @RoundtableSpace