GitHub radar
MiniMax H3 Turbo — create videos faster from images locally
With the help of ModelTC, MiniMax H3 can generate videos quickly: the Turbo version reduces the number of generation steps by several times with pre-installed ComfyUI workflows.
MiniMax, the creator of mainstream multimodal models, launched the H3 video generation model. It can automatically generate a video based on an image or a piece of text, along with a sound. With the help of ModelTC, MiniMax H3’s generation speed has been improved. This version only needs several times fewer generation steps to get the same result. And it comes with pre-installed ComfyUI workflows, so there is no need to manually set up a workflow. This model supports both image-to-video and text-to-video generation. In the first month, it was downloaded over 1.26 million times. You can try it online without installing at ModelTC’s official studio x2v.light-ai.top.
Why a vibe-coder should care
It’s a free offline service like subscription video generation. Compared to original MiniMax H3, ModelTC Turbo version is several times faster, but you don’t need a high configuration. If you have used ComfyUI to generate images, it will only take you several minutes to use it to generate videos.
How to install
Copy this and send it to your agent — Claude Code, Codex, any of them:
Install the video model from https://huggingface.co/lightx2v/Minimax-h3-Turbo via ComfyUI, generate a test clip from a photo I provide, and show me the result. Ask if you need any credentials or permissions.
A Mac with 32 GB of RAM or more, or a GPU with 24 GB of VRAM — video models like MiniMax H3 need serious hardware. You can try it without installing anything at the online studio: x2v.light-ai.top
Open on Hugging Face▌ More finds