---
id: 20260805-T0-04
title: "MiniMax-H3开源：全能多模态生成15秒带声视频"
title_en: "MiniMax-H3 Now Available on MLX: Generates 15-Second Video with Audio"
url: https://ai.daily.yangsir.net/daily/20260805-T0-04
issue_date: 2026-08-05
publish_date: 2026-08-04T19:10:09.000Z
category: release
source_name: "Simon Willison"
source_url: https://simonwillison.net/2026/Aug/4/minimax-h3-mlx/#atom-everything
---

# MiniMax-H3开源：全能多模态生成15秒带声视频

MiniMax发布全能多模态模型MiniMax-H3，支持文本、图像、音频和视频输入，并可生成最长15秒带音频的视频片段。该模型现已有Python包MiniMax-H3-MLX支持，可在Apple Silicon上运行。用户输入一段文字或图片即可快速生成短视频，适合快速原型制作、创意预览等场景。模型权重已开放，开发者可自由使用。

## English Version

**MiniMax-H3 Now Available on MLX: Generates 15-Second Video with Audio**

MiniMax released MiniMax-H3, a general-purpose, omni-modal generative system accepting text, images, audio, and video, generating up to 15-second video clips with audio. A Python package, minimax-h3-mlx, enables running on Apple Silicon. Users can input text or images to quickly produce short videos for prototyping or creative previews. Weights are open, so developers can use freely.

---

**来源**：[Simon Willison](https://simonwillison.net/2026/Aug/4/minimax-h3-mlx/#atom-everything)

**详情页**：https://ai.daily.yangsir.net/daily/20260805-T0-04

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*