comfyui-workflows/models/minimax/minimax-h3-javawock2609-r2v/README.md

MiniMax H3 R2V Javawock Pack

Summary

Same-day MiniMax H3 reference-to-video workflow from the javawock7618 pack. This is the strongest new character-consistency lane in the bundle and the best fit for practical local video reference work.

Model Family

• minimax

Status

• imported-only

Source

• Workflow pack page: https://huggingface.co/javawock7618/comfy-MiniMax-H3-workflows
• Workflow file: https://huggingface.co/javawock7618/comfy-MiniMax-H3-workflows/raw/main/MiniMax_int8_R2V-javano2609.1.json
• Original publisher: javawock7618
• Date imported: 2026-09-02

Developer notes

• The pack README describes this as the Reference-to-Video workflow.
• It is built around INT8 inference, Sol-Attn, SLA-Attention, Spectrum, Lightx2v LoRA support, optional Turbo-LoRA, native Extend support, Fun ControlNet Union V2V control, and latent upscaling.
• The workflow uses the pack's dedicated reference LoRA path for appearance retention.

Our notes

• Best new MiniMax H3 import today if you care about reference retention more than flashy motion.
• This is the right kind of update for the current watch area because it is practical control, not just a decorative LoRA.
• Still heavy on an RTX 3070, but the INT8 path is the main reason it is worth trying locally.

Required custom nodes

• ComfyUI-SolAttn_triton
• ComfyUI-Spectrum-MiniMax-H3
• ComfyUI-Switchboard
• ComfyUI-MiniMaxH3_LatentUpscaler
• ComfyUI-H3-FaceRefine

Required models

• Comfy-Org/MiniMax-H3
• minimax_h3_ref_lora_rank_256_bf16.safetensors
• minimax_h3_fun_controlnet_union_pruned_bf16.safetensors or minimax_h3_fun_controlnet_union_pruned_int8_convrot.safetensors if V2V control is used
• lightx2v/Minimax-h3-Turbo if the accelerated path is enabled
• drbaph/MiniMax-H3-Turbo-Lora-ComfyUI if the turbo path is enabled

Notes

• Best same-day MiniMax workflow for reference-guided video.