Skip to content

MiniMax H3

MiniMax's open-weight omni-modal video model, which generates 4 to 15 second clips with native stereo audio from text, images, video and audio. It is first on LMArena image-to-video, but its community license excludes the EU, UK, South Korea and the United States.

MiniMaxVideo modelReleased July 2026Data checked October 18, 2026

Where people rank it

#1

Turning images into video (tied), 57,112 votes

#4

Creating video (tied), 11,417 votes

From blind votes on LMArena, where people pick the better of two anonymous answers. See the full tables on the AI leaderboard.

Strengths and weaknesses

  • First on LMArena image-to-video, on more than 57,000 votes
  • Open weights on Hugging Face (about 4.4 million downloads)
  • Cheap API: $0.08 per second at 768p
  • The license does not allow use in the EU, UK, South Korea or the United States without a separate agreement
  • Lower on LMArena text-to-video than on image-to-video

Best for

  • Animating images into video
  • Self-hosting video generation outside the excluded regions

Every score we have

Each bar compares this model with the best result any model on this page has on the same test.

Compare with