MiniMaxLaunchAI ModelsAugust 26, 2026

MiniMax H3 open-weight video model tops Video Arena

MiniMax released H3, a 33B open-weight audio-video model, now the #1 open model in Video Arena for both text-to-video and image-to-video. It's live on fal, OpenRouter, Vercel AI Gateway, and Leonardo AI, with community tools emerging.

How this story unfolded

4 weeks · 0 reports · 16 community posts · 16 of 17 shown

Jul 31
Aug 3
Aug 4
- SITUATION DETECTED: MiniMax-H3 is now the #1 open model in Video Arena across both Text-to-Video...
Aug 5
- We just dropped SOTA open weights for our MiniMax‑H3 video model, and so much happened in 48 hours....
Aug 9
- H3 Motion Context v0.2.0 - reference mode support, and the visible seam at joins is fixed. New workflow included with both fl2va and ref2va in one workflow.
Aug 10
Aug 26
- MiniMax H3 is at @anyscalecompute’s #RaySummit today in SF. @VictorSuOrtiz will break down our 33B...
Aug 27
Aug 29
- Tip for Minimax H3 max quality

MiniMax by email

Get an email when MiniMax has news

No news that day, no email.

MiniMax H3 open-weight video model tops Video Arena

How this story unfolded

MiniMax by email

More stories today