Skip to content
Intelligence IAAug 21, 2026Intelligence IA
Article

MiniMax released open weights for H3, a model that fuses text, image, video, and audio as context and outputs native stereo sound...

Up to 15 seconds at 2K. Community benchmarks show the 768p Base model running on consumer GPUs in minutes.

Redaction Data Cube AISource: Pandaily
01

Brief source

MiniMax released open weights for H3, a model that fuses text, image, video, and audio as context and outputs native stereo sound up to 15 seconds at 2K. Community benchmarks show the 768p Base model running on consumer GPUs in minutes.