MiniMaxLaunchAI ModelsAugust 9, 2026

MiniMax drops SOTA open weights for its H3 video model

Within 48 hours of MiniMax open-sourcing H3, the community ran it on $280 gaming GPUs and offline MacBooks and built unplanned tooling, including a Mac inference engine from the creator of Redis. On ThursdAI, H3 was compared directly alongside Seedance, FLUX and WAN as the "leading" open video model.

How this story unfolded

11 days · 13 reports · 131 community posts · 144 of 150 shown

  1. Jul 31
  2. Aug 1
  3. Aug 2
  4. Aug 3
  5. Aug 4
  6. Aug 5
  7. Aug 6
  8. Aug 7
  9. Aug 8
  10. Aug 9
  11. Aug 10

MiniMax by email

Get an email when MiniMax has news

No news that day, no email.

More stories today

Open the live feed