MiniMax H3 is Insane — T2V, I2V, FFLF & Turbo LoRA Full Workflow in ComfyUI

In this video, I'll show you how to run MiniMax H3 locally in ComfyUI — one of the most advanced open-source video generation models right now, with native stereo audio, multi-shot consistency, and built-in text rendering. I'll walk you through T2V, I2V, First & Last Frame, and Text-to-Last-Frame workflows, plus how to use the H3 Turbo LoRA by Larryvrh to get high-quality results in just 6 steps.

Featuring MiniMax H3huggingface.co H3 Turbo LoRA by Larry VRhuggingface.co

What you'll learn

• Optimizing for 12GB VRAM — Run the 40GB H3 model using pruned int8 models
• T2V, I2V, First & Last Frame, Text-to-Last-Frame workflows
• H3 Turbo LoRA integration — high quality in 6 steps
• Native audio, sound effects & in-video text rendering
• Multi-shot & consistent character generation
• Resolution & aspect ratio control (480p to 1080p)
• 8GB VRAM alternative models

Workflow

Workflow downloadbuymeacoffee.com

Prompt guide-

MiniMax-AI/MiniMax-H3github.com

Model links

models/diffusion_models

INT8 Pruned (12GB)huggingface.co 8GB VRAM modelshuggingface.co

models/text_encoders

qwen3vl_32b_minimax_h3_nvfp4_awq.safetensorshuggingface.co

models/vae

minimax_h3_video_vae_fp16.safetensorshuggingface.co minimax_h3_audio_vae_fp32.safetensorshuggingface.co Preview VAE (memory efficient)huggingface.co

models/loras

H3 Turbo v4huggingface.co

models/custom_nodes

H3 Turbo Nodegithub.com

• 👍 Support the Channel If you found this helpful:

• 👍 Like the video
• 🔔 Subscribe for more AI workflows
• 💬 Comment what you want to see next

#MiniMaxH3#ComfyUI#AIVideo#GenerativeAI#TextToVideo#ImageToVideo#NativeAudio#OpenSourceAI