MiniMax Music 2.0

• 🎯 Full song generation with vocals and lyrics^1
• 📏 Provide your own structured verse/chorus lyrics^1
• 🔧 Style prompt plus lyrics drive the song^1
• 💬 Sings your supplied words into complete tracks^1
• 🆕 Builds on earlier MiniMax Music versions
• 🏢 Built by Chinese AI lab MiniMax
• 🌐 Available through the MiniMax music generation API^1

MiniMax

MiniMax is an AI company building generative models across multiple modalities, with a focus that spans both language understanding and audio creation. Their rapid release cadence in early 2026—delivering several new models within just a few months—reflects…

MiniMax Music 2.0 is a text-to-music model that generates complete songs — with vocals, lyrics, and instrumentation — rather than short loops or instrumental clips.^1 You provide your own lyrics structured with verse and chorus sections plus a style prompt describing the song, and the model sings the words back as a finished arrangement.^1 It is exposed through the MiniMax music generation API.^1

The model sits within MiniMax's music line, which the lab has continued to iterate. After 2.0, MiniMax released MiniMax Music 2.5 and MiniMax Music 2.6 as newer versions in the same family. Version 2.0 remains a lyrics-driven endpoint oriented around workflows where you supply your own words and section structure.

It is aimed at songwriters, content creators, and developers building audio applications who want a complete song from lyrics and a text prompt.^1 Because the style direction and the lyrical content are provided separately, the same words can be paired with different style prompts to explore varied renditions.^1