# MiniMax H3 Motion Enhancer

## Summary

Same-day `MiniMax H3` motion-focused LoRA plus its portable ComfyUI workflow. This is a practical 8-step lane when you want more animation energy from the MiniMax H3 family, not just another style-only drop.

## Model Family

- `minimax`

## Status

- `imported-only`

## Source

- Model page: <https://huggingface.co/rzgar/minimax-h3_fl2v_8Step_motion_enhancer>
- Workflow file: <https://huggingface.co/rzgar/minimax-h3_fl2v_8Step_motion_enhancer/resolve/main/workflow/minimax_h3_i2v_MoEn_workflow.json>
- Source README: <https://huggingface.co/rzgar/minimax-h3_fl2v_8Step_motion_enhancer/raw/main/README.md>

## Developer notes

- The source card frames this as an 8-step LoRA for `MiniMax-H3 FL2VS`.
- The README explicitly says to disable the vanilla `lightX2V` LoRA when using this one.
- The source README focuses on motion and anatomy quality, with an NSFW-leaning description.

## Our notes

- This looks like a real utility add if you already use MiniMax H3 and want more motion from an 8-step setup.
- It is more specialized than a general-purpose quality booster.
- On an `RTX 3070` with `8 GB` VRAM, expect it to feel tight rather than easy.

## Required custom nodes

- `comfyui-easy-use`
- MiniMax H3-compatible ComfyUI support

## Required models

- `Comfy-Org/MiniMax-H3`
- `Comfy-Org/MiniMax-H3` `minimax_h3_fl2va_pruned_int8_convrot.safetensors`
- `Comfy-Org/MiniMax-H3` `minimax_h3_video_vae_fp16.safetensors`
- `Comfy-Org/MiniMax-H3` `minimax_h3_audio_vae_fp32.safetensors`
- `Comfy-Org/MiniMax-H3` `qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors`
- `lightx2v/Minimax-h3-Turbo` `minimax_h3_fl2v_turbo_8step_v1.0_comfyui_bf16.safetensors`
- Optional alternate path from the workflow notes: `minimax_h3_fl2va_fp16attn_fp8.safetensors`
- LoRA: `minimax-h3_fl2v_8Step_motion_enhancer.safetensors`

## Notes

- Good same-day MiniMax H3 import for the current workflow set.