MiniMax H3 Showcase: Hyper-Structured Prompts for Precise Cinematic Control

lolo780 · reddit · 2026-08-11

A user demonstrated the capabilities of the MiniMax H3 video model using a highly structured prompt. By defining subjects, keyframes, retention analysis, and detailed camera trajectories (tracking, arcing, pedestaling), the prompt achieved precise control over character actions, dialogue, and complex cinematic shots.

Original post →

More from Multimodal

Multimodal channel →