Uploaded August 2026 | Updated November 2026, 4 hours ago
MiniMax H3 is a new open-source AI video model by Hailuo AI, that runs entirely on your own computer inside ComfyUI: free, local, no subscription. In this video I cover the full setup: which model files to download for your GPU, all three workflows (text-to-video, image-to-video, and reference-to-video), how to prompt properly. I spent a week testing it, so everything here comes from real generations, including where it falls short.
📝Full written guide: bit.ly/4zphMSb
👇Everything you need:
1️⃣ ComfyUI: bit.ly/4x7sVoI
2️⃣ Workflows: bit.ly/4xzM9Um
3️⃣ H3 Models: huggingface.co/Comfy-Org/MiniMax-H3
🎓 Get started with ComfyUI: youtu.be/txgJ6d0WEDM
🌐 Run H3 Online:
- Text to video: bit.ly/4bE9eg3
- Image to video: bit.ly/4z77GoK
- Ref to video: bit.ly/4q7qUGW
🆙 Video Upscaler:
- Topaz video: bit.ly/3t04Otl
- SeedVR2.5: youtu.be/9kEux1X_I-U
⏲ Chapters:
0:00 What's new with H3
1:04 Install ComfyUI
1:18 Local H3 vs API
1:44 Download Models
1:58 VRAM Requirements
3:24 H3 Text to video & How to Prompt
4:29 Speed test
5:31 H3 Image to video
7:14 H3 Reference to video
8:55 What's next?
Miniax H3 is a new open source omnimodel capable of generating 15-second video clips with native stereo audio in one run.
This breakdown explores how Miniax H3 handles complex inputs by combining images, video clips, and audio into a single generation process. If you are looking for new tools to streamline your creative workflow, this model offers a specific approach to high-quality video synthesis that differs from standard generative AI platforms.
We look at the technical capabilities of this open source omnimodel and how it manages to process multiple media types simultaneously. By analyzing the output quality of Miniax H3, you can determine if this tool fits your current production needs without the typical limitations of older video generation software.
Disclaimer: Some links in the description are affiliate links. If you make a purchase through them, I may earn a small commission at no extra cost to you.
Support me on Patreon:
bit.ly/2MW56A1
🎵 Where I get my Music:
bit.ly/3boTeyv
🔈 Join my Discord server:
bit.ly/3qixniz
Join me!
Instagram: instagram.com/JustMDMZ
Tiktok: tiktok.com/@justmdmz
Twitter: twitter.com/JustMDMZ
Facebook: facebook.com/medmehrez.BSS
Website: mdmz.co
#Minimax #Aivideo #comfyui
Who am I?
-----------------------------------------
My name is Mohamed Mehrez and I create videos around visual effects and filmmaking techniques. I currently focus on making tutorials in the areas of digital art, visual effects, and incorporating AI in creative projects.
MiniMax H3 is a new open-source AI video model by Hailuo AI, that runs entirely on your own computer inside ComfyUI: free, local, no subscription. In this video I cover the full setup: which model files to download for your GPU, all three workflows (text-to-video, image-to-video, and reference-to-video), how to prompt properly. I spent a week testing it, so everything here comes from real generations, including where it falls short.
📝Full written guide: bit.ly/4zphMSb
👇Everything you need:
1️⃣ ComfyUI: bit.ly/4x7sVoI
2️⃣ Workflows: bit.ly/4xzM9Um
3️⃣ H3 Models: huggingface.co/Comfy-Org/MiniMax-H3
🎓 Get started with ComfyUI: youtu.be/txgJ6d0WEDM
🌐 Run H3 Online:
- Text to video: bit.ly/4bE9eg3
- Image to video: bit.ly/4z77GoK
- Ref to video: bit.ly/4q7qUGW
🆙 Video Upscaler:
- Topaz video: bit.ly/3t04Otl
- SeedVR2.5: youtu.be/9kEux1X_I-U
⏲ Chapters:
0:00 What's new with H3
1:04 Install ComfyUI
1:18 Local H3 vs API
1:44 Download Models
1:58 VRAM Requirements
3:24 H3 Text to video & How to Prompt
4:29 Speed test
5:31 H3 Image to video
7:14 H3 Reference to video
8:55 What's next?
Miniax H3 is a new open source omnimodel capable of generating 15-second video clips with native stereo audio in one run.
This breakdown explores how Miniax H3 handles complex inputs by combining images, video clips, and audio into a single generation process. If you are looking for new tools to streamline your creative workflow, this model offers a specific approach to high-quality video synthesis that differs from standard generative AI platforms.
We look at the technical capabilities of this open source omnimodel and how it manages to process multiple media types simultaneously. By analyzing the output quality of Miniax H3, you can determine if this tool fits your current production needs without the typical limitations of older video generation software.
Disclaimer: Some links in the description are affiliate links. If you make a purchase through them, I may earn a small commission at no extra cost to you.
Support me on Patreon:
bit.ly/2MW56A1
🎵 Where I get my Music:
bit.ly/3boTeyv
🔈 Join my Discord server:
bit.ly/3qixniz
Join me!
Instagram: instagram.com/JustMDMZ
Tiktok: tiktok.com/@justmdmz
Twitter: twitter.com/JustMDMZ
Facebook: facebook.com/medmehrez.BSS
Website: mdmz.co
#Minimax #Aivideo #comfyui
Who am I?
-----------------------------------------
My name is Mohamed Mehrez and I create videos around visual effects and filmmaking techniques. I currently focus on making tutorials in the areas of digital art, visual effects, and incorporating AI in creative projects.

![No jumpscares at the end of this video, I promise 😌 #short
Tutorial: https://youtu.be/Yf2G9ye37ms
Prompt: [spooky, halloween, scary, fantasy, mystical, mist, scary, fantasy, featured on artstation, 8k] No jumpscares at the end of this video, I promise 😌 #short](https://i.ytimg.com/vi/eJufMsB6hN0/mqdefault.jpg)








