MiniMax Music 3

MiniMax Music 3

Repackaged MiniMax Music 3 model files for ComfyUI, with a Text to Music workflow for making audio from a written prompt. For ComfyUI users who generate music.

Made by Comfy-Org on Hugging Face. These notes are Mindket's, written from its model page and model card; the model's own files and license are on Hugging Face.

About

What it is: This repository holds MiniMax Music 3 model files repackaged by Comfy-Org so they load in ComfyUI. The original model comes from MiniMaxAI. The files include diffusion models, text encoders and a VAE, and the card links a Text to Music workflow.

Who it suits: ComfyUI users who want to make music from a written description, on their own computer. The card is short and gives file locations and one workflow, so it assumes you already know ComfyUI.

Title: Who uses it, and how
Industry | Who uses it | How they use it | Example
Music making | Hobby producers | Generate music ideas from text | Describe a mood and genre, then listen to the result
Video and content | Creators | Make background tracks for projects | Generate a short track to sit under a video
Game development | Indie developers | Prototype music for a scene | Try several prompts for a menu theme
AI research | ComfyUI tinkerers | Compare the different file precisions | Swap one model file for another and compare output

Skip it if you do not use ComfyUI, because the files are packaged for that program.

Where to use it:

• ComfyUI, on Windows, macOS or Linux
• The Text to Music workflow template the card links
• Windows, macOS or Linux for downloading the files, with Python

Use case | Who | What they get back
Text to music | Creators | Audio generated from your written prompt
Workflow testing | ComfyUI users | A ready-made graph to run and modify
Precision tests | Tinkerers | Output from different model file versions

How to set it up: You need ComfyUI installed, plus Python and pip to download. Then you copy each file into the matching folder inside ComfyUI's models directory.

• ComfyUI
• Python and pip
• Disk space for the model files

1. Install the Hugging Face hub library.
2. Download the repository.
3. Copy the model files into ComfyUI folders, then open the workflow.

pip install -U huggingface_hub
hf download Comfy-Org/MiniMax-Music-3

Title: Where each file goes
Folder | Files
diffusion_models | minimax_music3_dit_fp16, minimax_music3_dit_fp32, minimax_music3_dit_int8_convrot
text_encoders | minimax_music3_text_encoder_bf16, minimax_music3_text_encoder_pruned_bf16, minimax_music3_text_encoder_pruned_int8_convrot
vae | minimax_music3_dav

How to use it: Open the Text to Music workflow in ComfyUI, make sure it points at the files you placed, write a description of the music you want and run it. The card lists three diffusion files and three text encoders without saying which to pick, so follow the workflow's own settings.

Tool | What it does
ComfyUI | The program that runs the model through a node graph
Text to Music workflow | A ready-made graph for turning a prompt into audio
diffusion_models folder | Holds the main music model file
text_encoders folder | Holds the file that reads your prompt
vae folder | Holds the file that decodes the result into audio

Example: You open the Text to Music workflow and describe a calm piano piece with soft strings. You run the graph, and it returns an audio clip. The card shows no sample output, so judge the result yourself.

For newcomers and experts: If you are new, learn the basics of ComfyUI first, then place the files and open the workflow. Run it once with a short prompt before changing anything.

For people in the field, the repository repackages a diffusion-transformer music model with its text encoder and VAE as single files. Some files are fp16 or fp32, some are bf16, and some are int8 with the convrot label. The card does not explain the variants.

Download the repository
Copy files into ComfyUI model folders
Open the Text to Music workflow
Write a prompt and run it
Get an audio clip

Good to know:

• License: Apache 2.0
• Cost to run: none named, you use your own computer
• Data it reaches: it downloads files from Hugging Face, and the card names nothing else it touches
• Not for: use outside ComfyUI, since the card only describes that setup
• The card gives no benchmark scores, sample audio or hardware notes

On Hugging Face, 13 November 2026

• Library: diffusion-single-file
• Downloads in the last 30 days: 1,077,395
• Likes: 260
• License: apache-2.0
• Based on: MiniMaxAI/MiniMax-Music3
• Tags: diffusion-single-file, comfyui

Links

• Model: https://huggingface.co/Comfy-Org/MiniMax-Music-3
• Files: https://huggingface.co/Comfy-Org/MiniMax-Music-3/tree/main

Get the model

Download its files with Hugging Face's command line:

pip install -U huggingface_hub
hf download Comfy-Org/MiniMax-Music-3

Then follow the usage in its model card. Nothing in this kit runs by itself.