Beginner guide · ComfyUI · 2026

MiniMax H3 Tutorial — Get Started in 5 Minutes

Learn what MiniMax H3 is, which files you need, how to load the official ComfyUI workflows, and how to create your first text-to-video or image-to-video clip.

Start the quickstart Model files

Independent guide. Not affiliated with MiniMax, Hailuo AI, Hugging Face, or ComfyUI.

Fast path

MiniMax H3 quickstart

1

Update ComfyUI

Use a recent ComfyUI build so the MiniMax H3 nodes and workflow templates are available. Restart ComfyUI after updating.

2

Download weights

Download the diffusion model, Qwen3-VL text encoder, video VAE, and audio VAE from the Comfy-Org MiniMax H3 repository.

3

Load workflow

Import the official T2V or I2V workflow JSON. Beginners should start with T2V because it needs no source image.

4

Generate and iterate

Run one short test clip, check motion and audio, then improve the prompt before increasing resolution or duration.

Overview

What is MiniMax H3?

MiniMax H3 is an open omni-modal video generation model from the MiniMax / Hailuo video family. It is designed to understand mixed context such as text, image, video, and audio, then generate short video with native stereo audio in the same workflow.

For beginners, the most practical way to try MiniMax H3 is through ComfyUI. ComfyUI gives you visual workflow files for text-to-video, image-to-video, and reference-to-video generation, so you do not need to write inference code from scratch.

MiniMax H3 ComfyUI workflow interface

MiniMax H3 ComfyUI workflow interface: prompt, model loader, sampler, and video output nodes.

Installation

How to install MiniMax H3 in ComfyUI

1. Install or update ComfyUI

If you already use ComfyUI, update it first. If not, install ComfyUI from the official GitHub repository and confirm that it opens in your browser.

Open ComfyUI on GitHub

2. Download model files

Use the Comfy-Org repackaged model repository. It lists the correct folder layout for diffusion models, text encoders, and VAEs.

Open Comfy-Org/MiniMax-H3

3. Put files in the right folders

ComfyUI/models/diffusion_models/
ComfyUI/models/text_encoders/
ComfyUI/models/vae/

Do not rename the files unless the workflow explicitly tells you to. Wrong folder names are one of the most common beginner errors.

4. Start with lighter files

If you are not on a large GPU, start with pruned or quantized files such as INT8 diffusion and NVFP4/AWQ text encoder variants. Move to full precision only after the workflow runs successfully.

Official workflow templates

Download the workflow JSON files

T2V workflowText prompt to video I2V workflowImage plus prompt to video R2V workflowReference media to video

Example T2V prompt

A cinematic close-up of a glass tea cup on a rainy window sill, warm indoor lighting, slow camera push-in, soft reflections, natural steam movement, gentle ambient room tone.

Text-to-video

How to use MiniMax H3 T2V

Text-to-video is the easiest MiniMax H3 mode. Load the T2V workflow, write a prompt that describes the subject, motion, camera, environment, and audio mood, then queue a short generation.

  1. Open the T2V workflow JSON in ComfyUI.
  2. Confirm every model loader points to the files you downloaded.
  3. Start with a short clip and moderate resolution.
  4. Use concrete motion words: “slow push-in”, “handheld pan”, “hair moving in wind”.
  5. Only increase settings after the first test succeeds.

Image-to-video

How to use MiniMax H3 I2V

Image-to-video is best when you want character, product, or composition consistency. Use a clean starting image, then describe the motion you want instead of rewriting everything in the image.

  1. Load the I2V workflow JSON.
  2. Upload a clear source image with the right aspect ratio.
  3. Describe motion, camera movement, lighting changes, and audio.
  4. If faces or hands drift, reduce aggressive motion and try again.
  5. Save the workflow once all file paths are correct.

Example I2V prompt

Animate the portrait into a subtle talking-head video. The person smiles gently, blinks naturally, and turns slightly toward camera. Keep the face identity stable. Quiet studio ambience, soft voice tone.

Beginner mistakes

Common MiniMax H3 setup problems

Out of memory

Use lighter model variants, lower resolution, shorter duration, or a cloud GPU. Do not debug prompt quality before the workflow can run once.

Missing model error

Check folder names exactly: diffusion models, text encoders, and VAEs belong in different ComfyUI model folders.

Weak motion

Add explicit verbs and camera language. “A woman in a cafe” is weaker than “a woman lifts a cup while the camera slowly pushes in”.

Audio mismatch

Describe sound in the same prompt: ambience, dialogue tone, music style, and sound effects. Keep the audio request simple at first.

FAQ

MiniMax H3 frequently asked questions

How much VRAM do I need for MiniMax H3?

MiniMax H3 is a large video model, so hardware requirements depend on file variant, resolution, duration, and workflow. Beginners on consumer GPUs should try pruned or quantized files first. If you still hit out-of-memory errors, use a cloud GPU or lower your settings.

Can I use MiniMax H3 for free?

You can download open model files under the stated community license and run them on your own hardware. Free still means you pay with local GPU hardware, electricity, storage, or cloud GPU rental if your machine is not powerful enough.

Does MiniMax H3 support text-to-video and image-to-video?

Yes. The Comfy-Org workflow templates include local T2V and I2V workflows, plus an R2V reference workflow. Start with T2V for the simplest first test.

What formats should I use?

Use common image formats such as PNG or JPG for input images. Exported video format depends on the ComfyUI workflow and video combine node settings, but MP4 is usually the most convenient format for sharing.

Can MiniMax H3 generate 2K video locally?

Base local workflows commonly start at lower output sizes for practical inference. Some 2K regeneration features may be hosted or workflow-specific. Check the latest official MiniMax and ComfyUI documentation before promising 2K local output to clients.

Does MiniMax H3 support audio generation?

Yes. MiniMax H3 workflows can generate video with native stereo audio, including simple dialogue tone, ambience, music, and sound effects described in the prompt. For the best first result, keep the audio request short and concrete.

How to fix "model not found" error in MiniMax H3?

A model not found error usually means the file is missing, renamed, or placed in the wrong ComfyUI folder. Check that diffusion models are in models/diffusion_models, text encoders are in models/text_encoders, and VAEs are in models/vae, then restart ComfyUI and reload the workflow.

Where can I download the MiniMax H3 workflow JSON?

Download the official workflow JSON files from the Comfy-Org workflow_templates repository: T2V, I2V, and R2V.

What is the difference between MiniMax H3 and Hailuo?

MiniMax is the company and Hailuo is its video product line or service branding. MiniMax H3 refers to the H3 video generation model release that can be used through supported ComfyUI workflows and related hosted services.

Can MiniMax H3 run on RTX 3060 / 4090?

An RTX 4090 with 24GB VRAM is a much more realistic local starting point when using pruned or quantized files. An RTX 3060 12GB may be too tight for comfortable local MiniMax H3 generation unless you use very light settings, aggressive quantization, CPU offload, or cloud execution.

Next step

Install ComfyUI, download the model files, load the T2V workflow, and generate one short test clip. After that, improve the page by replacing placeholders with real screenshots from your run.

Back to installation