Skip to content
Music model · MiniMax

MiniMax Music 2.6 when the ad needs a chorus

An instrumental bed can be bought anywhere. A jingle that sings your own line in Polish and sounds recorded rather than translated cannot. This engine takes a lyric sheet with [Verse] and [Chorus] tags, arranges around them, and its diction is clean enough that the words land on the first listen.

Rendered by this model

Turn the sound on and listen

–:–– / –:––
A bakery jingle, our words: "Świeży chleb każdego rana… Wpadnij do nas, jest otwarte" — sung in Polish.This was generated by the model this page is about, exactly the way the app runs it. It is not a photo shoot and not client work.

Specification

Built by
MiniMax
Starts from
a written brief
Track length
the engine decides
Your own lyrics
optional, with [Verse] and [Chorus] tags
Vocals
a switch: sung or instrumental
Render time
about 150s
Cost
1 credit per track

What it is good for

Polish diction we verified
The track beside this was run back through speech recognition: the detected language is Polish and the transcript returns the exact lines we wrote. It is the only way to judge material like this beyond "sounds fine".
Verse and chorus where you put them
[Verse], [Chorus] and [Bridge] in the sheet are not decoration — the engine builds the arrangement around them, so the hook lands where the ad needs it.
Or it does not sing at all
It has a real instrumental switch rather than a request in the brief. The same sheet gives you a sung jingle one run and a bare bed for a voiceover the next.

When to pick something else

The longest wait of the five: around two and a half minutes, ten times what Lyria 3 takes. If you are looking for a mood rather than words this is the wrong pick — Lyria 3 or Lyria 3 Pro for a bed, ElevenLabs Music when the length has to be exact. It takes no picture as a mood hint either; it starts from the brief and the sheet only. And you do not choose the melody: tempo, genre and voice type go in the brief, but the tune is the engine’s call, so a jingle that has to stay the same forever will take a few attempts.

See every model

How to brief this model

The things this engine handles differently from the rest. The habits that hold for every model are collected separately, in the help centre.

  1. Mark the sections out with tags. A verse and a chorus tagged separately produce an arrangement where the chorus actually returns rather than dissolving mid-track.
  2. Write words that can be sung. Short words and open syllables come out clearly here; a pile-up of consonants turns into a sound that resembles a word without being one.
  3. Listen to every line before you publish the track. We checked our own sample with speech recognition, and that is the only reason we know the Polish words are intelligible.
  4. Do not expect a repeatable melody. Tempo, genre and the kind of voice you can describe in the brief, but the melodic line is the engine’s decision — for a jingle that has to be the same every time, expect several attempts.
  5. Plan for the wait: about two and a half minutes, ten times longer than Lyria 3. If you are after a mood rather than words, this is the wrong choice.
How to write a brief that works

Questions about this model

Try this model on your own product

Sign up without a card and get 40 credits. Switching model is one click, so you can send the same brief through several of them and keep the best frame.