FrontierThe story, in brief
MiniMax Releases MiniMax-Music3: An Open-Weights Music Model Generating Complete Five-Minute Songs From Lyrics and a Structured Caption
MiniMax drops open-weights music generation that goes the distance: full five-minute songs in one pass, not loops.
The KeyNews take
Why it matters
A capable open-weight model for music generation shifts the competitive landscape away from closed APIs. Practitioners building music features now have a viable alternative to proprietary services; enthusiasts tracking the model race see text-to-audio catching up to text-to-image in capability and accessibility.
The key facts
6 to knowOpen-weights release (not gated)
Text-to-music model with lyrics + structured caption input
Five-minute song generation in single pass
32 kHz, 16-bit stereo WAV output
Architecture and serving paths disclosed
License conditions specified (actionable for deployment)
Go to the source
MarkTechPostmarktechpost.com
Publisher excerpt: MiniMax released MiniMax-Music3, an open-weights text-to-music model. Given lyrics with section tags and a structured caption, it generates a complete song of up to five minutes in a single pass, as 32 kHz, 16-bit stereo WAV. Here is the architecture, the three serving paths, and the license…