MiniMax-H3 Released: Omni-modal System Generates 2K Video with Native Stereo Audio
VoidAsuka · x · 2026-08-03
MiniMax has publicly released MiniMax-H3, a general-purpose omni-modal generative system, on Hugging Face. The model offers unified understanding of multimodal contexts—including text, images, video, and audio—and can generate videos up to 2K resolution and 15 seconds in length with native stereo sound.
Designed with task generalization in mind, H3 possesses broad multimodal comprehension and generation capabilities straight out of the pre-training stage, enabling it to follow complex instructions effectively. It supports various input modes like first-and-last-frame control and handles 11 languages, including Chinese and English.
Related event: MiniMax open-sources 33B multimodal video model H3(35 posts)→
More from Multimodal
- Fan animation made in Blender with Claude Code (Opus 5.5) — Icy-Appointment3583 · 2026-10-16
- Making Blender animations with Opus, Astra and Argon: how you prompt still matters — chongdashu · 2026-10-16
- Tessellation-free geometric queries on displaced surfaces cut caustics variance 10x — ssh4net · 2026-10-16
- Complex multi-link CAD mechanism finally animates, dev celebrates the breakthrough — zetalyrae · 2026-10-16
- OpenMOSS releases MOSS-Image-SFT: Apache 2.0 text-to-image model on Hugging Face — bdsqlsz · 2026-10-16
- Hyperflow open-sourced: 49 sampling steps down to 8, ~3x faster video generation — mhdfaran · 2026-10-16