All the models in one place
Run Generate Music (MiniMax v2.6) and every other model on Glif from one chat. One account, one balance, no API keys to wire up.
Generate Music (MiniMax v2.6)
Head-to-head ratings
Blind votes by the Glif team, as of
- Text-to-music rating
- Needs more ratings (4 of 20 games)
Elo rating centred on 1500, compared only within this mode.
Show score details
31.5% predicted preference against a 1500-rated reference · Wins 1 / Losses 3 / Ties 0.
Bradley–Terry ratings center on 1500 and have no maximum. Preference = 100 / (1 + 10^((1500 − rating) / 400)). LLM-judge scores are not included in this export.
Test outputs
Compare all modelsGenerate Music (MiniMax v2.6) in practice
What it does and when to use it.
Model summary
MiniMax Music 2.6 creates songs or instrumental tracks from a text description. Describe the genre, instruments, mood, and vocal style. You can provide lyrics or have the model write them from your prompt.
Capabilities
Text to music
Music generation
Song generation
Prompt conditioned audio
Vocal music generation
Lyric conditioned generation
Style controlled generation
Suggested use cases
- Generate original songs from detailed text prompts
- Create vocal-forward tracks (e.g., rock/grunge) with specified mood and structure
- Rapid prototyping of musical ideas (genre/tempo/instrumentation/arrangement)
- Custom lyric-driven songs when you provide the lyrics
- Soundtrack-style music generation by specifying scene, pacing, and production traits