Bullish

MiniMax H3 Open Source Launches with 2K Video and Native Audio

2026-08-03 10:50:56

MiniMax releases open-source universal video model H3, supporting multimodal input and generating 2K resolution, 15-second clips with native stereo audio.

Woofun AI reports that MiniMax has officially open-sourced its new generation universal video model, MiniMax H3. The model is designed to uniformly process multimodal contexts comprising text, images, videos, and audio. It enables the generation of video content with up to 2K resolution, durations of up to 15 seconds, and native stereo audio capabilities.

WOOFUN AI

Impact Assessment · Quick Read

Open-sourcing MiniMax H3 lowers barriers for developers to integrate high-fidelity video generation into their applications. The inclusion of native stereo audio and 2K resolution sets a new baseline for open-weight models, potentially accelerating competition in the generative media sector.
Generated by WOOFUN AI · For reference only, not investment advice

Comments

Me
Replying to @User
0/800

No comments yet.

Notifications

Sign in to view messages
View all messagesManage subscriptions