MiniMax opens H3, 33B video model with audio, on Hugging Face

MiniMax opened the weights of H3, a 33B-param video model that handles text-to-video, image-to-video and reference-to-video with audio, ready for consumer GPUs via diffusers and ComfyUI. Within four days the community shipped a distillation LoRA cutting sampling from 20 to 4–8 steps, with H3 also running on Macs.
How this story unfolded
9 days · 2 reports · 120 community posts · 122 of 131 shown
- Jul 31
- Aug 1
- Aug 2
- Aug 3
- Aug 4
- Aug 5
- Aug 6
- Aug 7
- Aug 8
MiniMax by email
Get an email when MiniMax has news
No news that day, no email.
More stories today
- Paper examines the limitations of current AI evaluation methods
- OnlyHuman filter list removes AI-generated SEO spam from search results
- Qwen tokenizes 330-line code into 1,609 tokens; Gemma needs 4,258
- LifeOS: open-source AI harness for personal growth and work
- MINIMAX video drops Indiana Jones into Mortal Kombat