MiniMax releases open weights for H3 video generation model
MiniMax-H3 is now the top-ranked open model in Video Arena for both text-to-video and image-to-video generation. Community developers have already created distillation LoRAs that reduce sampling steps from 20 down to 4–8, enabling performance on consumer hardware like the RTX 5090.
How this story unfolded
11 days · 10 reports · 136 community posts · 146 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
- Koharu translates manga pages locally using OCR, inpainting, and LLMs
- Orchestration tool runs AI coding agents in parallel, compares answers
- Real-time Gaussian splats generated on mobile devices
- Newtake AI shows off completely AI-generated rap music video
- ChatGPT Work used to install OpenClaw and Ollama, run local model