MiniMax releases H3, SOTA open-weights video model
H3 is now the #1 open model in Video Arena for both text- and image-to-video, tied #3 overall, and SOTA on Artificial Analysis. Within 48 hours the community added native ComfyUI support, a distillation LoRA cutting sampling from 20 to 4–8 steps, and runs on $280 GPUs, an RTX 5090, and offline MacBooks.
How this story unfolded
11 days · 11 reports · 134 community posts · 145 of 150 shown
- Jul 30
- Jul 31
- Aug 1
- Aug 2
- Aug 3
- Aug 4
- Aug 5
- Aug 6
- Aug 7
- Aug 8
- Aug 9
- Aug 10
MiniMax by email
Get an email when MiniMax has news
No news that day, no email.
More stories today
- Scoble envisions AI agents mapping 3D printing businesses
- Tool enables AI agents to build fullstack apps from prompts
- Tencent elevates WorkBuddy as a top strategic AI priority
- Apple denies Qwen integration launched in China after guide pulled
- Depth Any Panoramas creates depth maps for panoramic imagery