# MiniMax Open-Sources Music Model MiniMax-Music3

- Category: Flash (快讯)
- Published: 2026-08-14T11:56:00.000Z
- Source date: 2026-08-14
- HTML: https://www.88z.ai/en/news/flash-20260814-9dd1e336e4c7fa2b
- Markdown: https://www.88z.ai/en/news/flash-20260814-9dd1e336e4c7fa2b.md
- Other language: https://www.88z.ai/news/flash-20260814-9dd1e336e4c7fa2b.md

MiniMax has released the open-source music model MiniMax-Music3, capable of generating lyricized songs up to five minutes long. Using a hierarchical architecture with an 8B global model and a 0.6B local model, it outputs 32kHz stereo WAV audio while maintaining consistency in theme, rhythm, and vocal identity across full song structures.

## At a glance

- Uses a hierarchical architecture with an 8B global model and a 0.6B local model
- Supports generating 5-minute 32kHz stereo songs while maintaining structural coherence

## The update

MiniMax has officially open-sourced its latest music generation model, MiniMax-Music3, which is capable of producing songs with lyrics lasting up to five minutes. Technically, the model employs a hierarchical design strategy: an 8-billion parameter global model handles long-range semantic understanding and overall song structure, while a 0.6-billion parameter local model restores fine-grained acoustic information to ensure audio detail. The final output is a 32kHz stereo WAV file. The model effectively maintains thematic coherence, rhythmic stability, and vocal identity consistency during generation, fully presenting standard song arrangement progressions from intro and verse to chorus.

## Sources

- [IT之家](https://www.ithome.com/0/989/722.htm) (2026-08-14)

This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.