MiniMax H3 LoRA in ComfyUI: Fast Action + Hybrid Trick @ComfyUIworkflows
MiniMax H3 LoRA in ComfyUI: Fast Action + Hybrid Trick  @ComfyUIworkflows
Uploaded September 2026 | Updated November 2026, 1 day ago
Alibaba PAI released an 8-step acceleration LoRA for MiniMax H3. I tested it in
ComfyUI against the 4-step turbo LoRA I have been using, on the same scenes and the
same prompts.

In the video:

Fast motion and a fighting scene on both LoRAs, same prompt
What happens to sound on 4 steps
A prompt the 4-step result ignores
One type of scene where this model still fails
Generation times on a 5090, and how I cut mine down using both LoRAs in one workflow

Timings, settings and the full node setup are shown on screen.

WHERE THE FILE GOES
This one does not go into your loras folder. It goes into ComfyUI/models/pdd_acc/.
If that folder is missing, create it with that exact name.

Two files were released, one for FL2VA and one for Ref2VA. Match the file to the
model you are running. The wrong file gives a wrong result and it will look like the
LoRA is broken.

LINKS
Custom node: github.com/Jalen-Brunson/ComfyUI-MiniMax-H3-PDD-Acc
Official LoRAs from Alibaba PAI: huggingface.co/alibaba-pai/MiniMax-H3-Acc-LoRAs/tree/main
4-Step Turbo LoRA source: huggingface.co/lightx2v/Minimax-h3-Turbo/tree/main
Workflow: patreon.com/ComfyuiWorkflowBlog/posts/minimax-h3-8-my-168587999


#MiniMaxH3 #ComfyUI #AIVideo
MiniMax H3 LoRA in ComfyUI: Fast Action + Hybrid TrickQwen Image Edit in ComfyUI: GGUF & FP8 Low-VRAM WorkflowMiniMax H3 ComfyUI Workflow: 3× Faster, Ref2VA & Turbo LoRAWAN 2.2 Animate in ComfyUI: InfiniteTalk + Uni3C — Low VRAM (GGUF Q4) [Free Workflow]ComfyUI Pose Hack: Steal ANY Pose using Qwen-VL & Z-Image (No VNCCS)ComfyUI Flux Plus IterComp Model Workflow | Refining and Upscaling | Enhanced | TutorialsFLUX 3 Image Klein in ComfyUI: Full Tutorial & WorkflowText-to-Video in 6 Seconds! Fastest Video AI Model | ComfyUI LTX Model Workflow Advanced All in OneFlux 2 Klein 9B ComfyUI: 100% Perfect Face & Pose SwapsInPainting Like Never Before: ComfyUI Advanced Flux InPainting Workflow, Flux fill & ReduxFIBO 2.0 ComfyUI Tutorial: Dev vs Turbo, VGL & Image EditingJoyAI ComfyUI: Edit Multiple Images at Once (+ Free Workflow)
ComfyUI Workflow Blog |

MiniMax H3 LoRA in ComfyUI: Fast Action + Hybrid Trick

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER