MiniMax H3 Workflows (T2V, I2V, First-Last-Frame, Ref2V, Lipsync) with SageAttention — v1.1 (faster+upgraded) — v1.6 (betterquality fl2v)

This MiniMax H3 ComfyUI workflow set makes video from text prompts, images, or reference video; its variants are labeled for text-to-video, image-to-video, first/last-frame, reference-to-video, and lipsync.

The archive contains 12 JSON workflow files, including prompt generators, I2V workflows, and Ref2V workflows; filenames distinguish prompt-enhancer and Larryvrh-node variants.

The extracted I2V workflows include MiniMaxH3ImageToVideo, while Ref2V workflows include MiniMaxH3ReferenceToVideo; video paths include VHS_VideoCombine.

Prompt-generator files contain multiline text controls; image-to-video files include LoadImage, while reference-to-video files include VHS_LoadVideo and VHS_LoadAudioUpload.

The creator links to MiniMax H3 model files, Heretic text encoders in INT8 and NVFP4, and a ComfyUI build.

Listed node packs include comfyui_steudio, comfyui-kjnodes, comfyui-lora-manager, comfyui-minimax-h3-turbo, comfyui-pixaroma, and rgthree-comfy.

Other listed node packs are comfy-image-saver, comfyui-easy-use, comfyui-rmbg, comfyui-videohelpersuite, comfyui_nvidia_rtx_nodes, and comfyui-spectrum-minimax-h3.

Model filenames listed across variants include gemma-4-E4B-it-ultra-uncensored-heretic-fp8.safetensors, Heretical-Qwen3.5-9B-fp8.safetensors, minimax_h3_audio_vae_fp32.safetensors, minimax_h3_fl2va_int8_convrot.safetensors, minimax_h3_turbo_v4_step600_ema.safetensors, minimax_h3_

Choose a Base Prompt generator for prompt generation, an I2V file for image-to-video, or a Ref2V file for reference-to-video.

Sugerencia · sin verificar

For the prompt-enhanced or Larryvrh-node options, choose a file whose name includes that variant label.

Sugerencia · sin verificar

Check the selected file’s listed node packs and model files before loading, since the 12 variants do not all list the same requirements.

Sugerencia · sin verificar