Skip to content
Pricing UnlimitedLog inSign up free

Prompting guide

How to prompt MiniMax Music 3

How to prompt MiniMax Music 3: describe the genre, voice, and arrangement, then write tagged lyrics it sings, with full songs and instrumentals to copy.

Try MiniMax Music 3MiniMax Music 3 model page

Quick start

Your first great MiniMax Music 3 prompt

Everything a beginner needs in one minute: what to know first, the rules that matter most, a template to fill in, and prompts to copy.

What to know first

MiniMax Music 3 takes two inputs with separate jobs. The prompt describes the music: the genre, the tempo and key, the mood and how it builds, the singer, and the instruments section by section. The Lyrics setting holds the words to sing, with section tags such as [verse] and [chorus] alone on their own lines. The description is never sung, and the lyrics never set the style, so keep each to its job. Name the singer in every sung song, or it can come out instrumental. Leave Lyrics empty or turn on Instrumental for no vocals, and pick a Length long enough for your lyrics: it is a limit, and the song often ends sooner.

The rules that matter most

  1. Describe the music, not the words

    The prompt is a brief to a band: the genre and subgenre, the tempo in BPM, the key, the mood from start to finish, and what the instruments do. The words to sing go in Lyrics, never in the prompt.

  2. Always name the singer

    Say who sings and how: "warm female lead, close and breathy in the verses, soaring in the chorus, soft stacked harmonies". A sung song whose description leaves out the voice can come out as an instrumental.

  3. Put each section tag on its own line

    In Lyrics, write [verse], [chorus], [bridge], and the other tags alone on their lines, with the sung lines under them. Words on the same line as a tag are dropped without a warning.

  4. Lay out the arrangement in time

    Say what plays in the intro, what enters at the first chorus, what drops out in the bridge, and how it ends. Without a plan, the model keeps the same band from start to finish.

  5. Size the lyrics to the Length

    A verse and a chorus fill about a minute; two verses, three choruses, and a bridge fill three or four. The song stops at the Length even mid-line, and ends on its own when the lyrics are done.

Fill in this template

Replace each part in square brackets with your own idea, and leave out a part you do not need.

Template
[Genre and subgenre], [tempo] BPM, [key]. [The mood, and how it builds]. Vocals: [who sings, and how]. Arrangement: [what plays first]; [what enters later]; [how it ends].
The template, filled in
Indie folk pop ballad, 84 BPM, G major. Tender and hopeful, opening hushed and swelling into a wide, glowing final chorus. Vocals: warm female lead, close and breathy in the verses, opening up with soft stacked harmonies in the chorus. Arrangement: fingerpicked acoustic guitar and felt piano carry the verses; cello and brushed drums enter at the first chorus; a gentle string swell lifts the last chorus, then the song falls back to guitar alone.
  • [Genre and subgenre] The style, as specific as you can: indie folk pop ballad, boom bap hip-hop, modern country with Americana roots.
  • [tempo] The speed in beats per minute: about 70 for a slow ballad, 95 for a mid-tempo groove, 120 to 140 for dance and rock.
  • [key] The key and scale, such as G major or E minor. Minor keys lean darker, major keys brighter.
  • [The mood, and how it builds] The feeling and its arc over the song: tender and hopeful, opening hushed and swelling into a wide final chorus.
  • [who sings, and how] The voice, its timbre, and its delivery in the verses and the chorus, plus harmonies: a raspy male lead, restrained in the verses and belting in the chorus.
  • [what plays first] The instruments that open the song: fingerpicked acoustic guitar and felt piano.
  • [what enters later] What joins and when: cello and brushed drums enter at the first chorus, a string swell lifts the last one.
  • [how it ends] The outro: the song falls back to guitar alone, rings out on one chord, or fades on the hook.

Prompts to start from

  • A synthwave song with a band brief

    Synthwave pop, 108 BPM, A minor. Driving and nostalgic, night-drive energy that builds to a euphoric chorus. Vocals: clear male tenor lead with a touch of reverb, doubled in the chorus, airy falsetto ad-libs at the end. Arrangement: pulsing analog bass arpeggio and gated-reverb drum machine from the start; shimmering pads and a bright lead synth hook in the chorus; a saxophone-like synth solo in the bridge; big 80s snare and chorus-drenched guitar in the final chorus.

    Why it works: Genre, tempo, key, and mood come first, the singer is cast plainly, and the arrangement says what enters at each section, including a [solo] for the synth. Pair it with tagged lyrics.

  • A rap verse with a sung hook

    Boom bap hip-hop, 90 BPM, D minor. Confident and reflective, head-nodding groove that never rushes. Vocals: male rapper with a warm, gravelly baritone and crisp, rhythmic delivery in the verses; a soulful female vocal sings the hook. Arrangement: dusty vinyl-sampled piano chops and a deep upright bass loop, punchy kick and snare with swung hi-hats, record scratches between sections, a muted horn stab in the hook.

    Why it works: Two voices get two jobs: the rapper takes the [verse] sections and the singer takes the [chorus], so the lyrics only need the tags.

  • A country song that rises

    Modern country with Americana roots, 96 BPM, D major. Warm, sunlit, and nostalgic, rising to a big singalong chorus. Vocals: male lead with a gentle Southern twang and an easy, storytelling delivery; a female harmony joins in the chorus. Arrangement: strummed acoustic guitar and harmonica open the song; pedal steel and a steady kick-snare groove enter in the first chorus; banjo rolls in the second verse; a fiddle line soars through the last chorus.

    Why it works: Each section adds one instrument, so the song keeps growing instead of starting at full strength.

  • A cinematic instrumental

    Epic cinematic orchestral score, 72 BPM, D minor to D major. Begins mysterious and fragile and grows into a triumphant, emotional climax. Arrangement: a lone solo cello over soft sustained strings and distant piano; French horns introduce the main theme; taiko drums and low brass build tension; a full orchestra with choir 'ah' pads bursts into the major-key theme at the climax; it resolves on a quiet piano and harp coda. Wide concert-hall reverb.

    Why it works: With Instrumental on and Lyrics empty, the whole prompt can go to the arrangement: a key change, the instruments in the order they enter, and the room.

  • A lo-fi study beat

    Lo-fi hip-hop beat, 80 BPM, F major. Mellow, cozy, and steady, for late-night studying. Arrangement: dusty Rhodes electric piano chords with a lazy swing, warm upright bass, soft boom-bap drums with brushed snare and vinyl crackle, a gentle muted trumpet melody in the second half, light rain ambience throughout. Tape-saturated, slightly detuned, intimate mix.

    Why it works: A purpose ("for late-night studying") sets the energy, and the production words at the end set the sound of the record. Turn on Instrumental.

Advanced

Advanced MiniMax Music 3 prompting

Techniques for more control once the basics work.

Structure: the three-part description

A short description works: "A warm acoustic pop song with intimate female vocals, fingerpicked guitar, and a gradual build into a wide final chorus." But a vague description gets a generic arrangement. For control, MiniMax recommends a structured description in three parts, the way its own prompt rewriter writes them: global metadata, vocal details, and the arrangement.

Global metadata is the genre and subgenre, the tempo, the key and scale, the emotional arc across the song, a scene it would play in, and the production profile: how wide, warm, and loud the mix is. Vocal details are the singer: gender, timbre, register, the delivery in the verses and in the chorus, harmonies and backing vocals, and vocal effects. The arrangement is a timeline: what enters, leaves, and changes in each section, and where the energy peaks.

Labeled lines work as well as a paragraph; headings, bullets, and bold text are removed before the model reads them. Decide every part yourself: whatever you leave open, the model fills with its most average choice.

  • Write the arc, not one mood: "starts restrained and builds to an explosive final chorus" tells the model how to shape the whole song.
  • Give a scene when it helps: "a smoky blues club late at night" or "a road trip at sunset" pulls the tone together.
  • Describe the mix: "warm and mid-range heavy, the lead guitar slightly off-center, relatively uncompressed".
  • Keep the description to a few hundred words for a long song: the description, the lyrics, and every moment of audio share one budget, so a very long text can end a five-minute song early.

A structured description

Global Metadata: Electric blues rock, 92 BPM, E minor. A confident, gritty swagger from the first bar, restrained storytelling verses that release into explosive guitar breaks, ending on a frenetic, improvised outro. A smoky, dimly lit club late at night. Live, organic production with a moderately wide soundstage, warm and mid-range heavy, relatively uncompressed.
Vocal Details: Male lead with a deep, gravelly baritone, conversational and close to spoken in the verses, soulful and melodic in the refrains, with shouted ad-libs during the solos. Moderate plate reverb and a subtle slapback delay.
Arrangement: A slightly overdriven electric guitar leads from intro to outro, trading comping and searing lead lines. A walking electric bass and a shuffle drum beat lock together throughout. A Hammond-style organ enters quietly in the verses and swells in the choruses and solos. The drummer adds fills and crash accents as the solos build.

Why it works: Each of the three parts does one job, and the arrangement follows every instrument through the song instead of listing them once.

Lyrics and section tags

The Lyrics setting holds the words to sing, one sung line per line. Section tags shape the song: [intro], [verse], [pre-chorus], [chorus], [post-chorus], [bridge], [instrumental], [solo], and [outro]. Each tag goes alone on its line, directly above the lines it covers. Words on the same line as a tag are dropped, and nothing warns you.

An instrumental passage is a tag with nothing under it: an [intro] before the first verse, a [solo] or an [instrumental] in the middle, an [outro] at the end. Describe what plays there in the arrangement part of the prompt so the two agree.

Write original lyrics only. A tag can also carry a cue for its section, such as [bass-quartet-rumbles-in] in MiniMax's own example; say the same thing in the arrangement.

  • Repeat the chorus in full every time it comes back; the model sings what is written.
  • Short, rhythmic lines are easier to sing than long sentences.
  • Tags work in any case: [Chorus] and [chorus] are read the same.

A tagged lyric sheet

[verse]
We wrote our wishes on paper lanterns
Down by the river when the night was new
Your hand was shaking as you lit the candle
And every word I wrote was about you
[pre-chorus]
The water carried all the light away
But I still see it when I close my eyes
[chorus]
So let them rise, let them rise
Little fires in a velvet sky
If the wind takes every one we made
I will hold the last one, I will light the way
[outro]
I will light the way

Why it works: This goes in the Lyrics setting, not the prompt: every tag sits alone on its line, and the sung lines follow it. Pair it with the indie folk description in the template above.

A song with an instrumental break

[verse]
They told me keep your head down, keep your voice low
Stay inside the lines and never let it show
[pre-chorus]
So light it up, light it up
[chorus]
We are the ones who never fell asleep
We are the fire that you couldn't keep
[bridge]
[solo]
[chorus]
We are the ones who never fell asleep
We are the fire that you couldn't keep

Why it works: The empty [bridge] and [solo] tags leave room for the guitar solo the description asks for, before the last chorus returns.

Vocals and language

Describe the singer as a casting note: gender, timbre, and register ("a deep, gravelly baritone", "a soft, clear soprano"), the delivery in each section ("conversational in the verses, belted in the chorus"), harmonies and backing vocals, and any vocal effects. For two voices, give each one its sections: a rapper in the verses, a singer on the hook.

The model sings the language of the lyrics. MiniMax demonstrates English and Mandarin Chinese; for another language, write the lyrics in it and name the language and the accent in the vocal details, and try a short Length first.

  • Without a voice in the description, a sung song can come out instrumental: always name one.
  • Vocal effects are words too: "light plate reverb", "a subtle slapback delay", "doubled in the chorus".

A Mandarin ballad

Mandopop ballad, 76 BPM, E-flat major. Bittersweet and tender, building to an emotional final chorus. Vocals: soft, clear female lead sung in Mandarin Chinese, intimate in the verses and soaring in the chorus, with light layered harmonies. Arrangement: solo piano opens; strings and a soft drum groove enter in the chorus; an electric guitar melody answers the vocal in the bridge; the last chorus lifts with full strings.

Why it works: The description stays in English and names the language of the singing; the lyrics themselves are written in Mandarin.

Length and how songs end

Length is the longest the song may run. The model ends the song on its own when the lyrics and the arrangement are done, so a short lyric sheet with a long Length gives a short song, and a long lyric sheet with a short Length is cut off at the limit, even mid-line.

Plan the lyrics for the Length: about a verse and a chorus for a minute, and two verses, three choruses, and a bridge for three or four minutes. To try a description quickly, use a short Length and a verse and a chorus, then render the full song once the sound is right.

  • Write an [outro] so the song ends instead of stopping.
  • Seed repeats a take: the same prompt, lyrics, Length, and seed give the same song again, and a new seed gives another performance of it.

LoRAs

MiniMax Music 3 LoRAs change the sound: a genre, a production style, or a vocal character. Import one from Hugging Face or CivitAI on the Import page, then pick it under Model on the Advanced page.

Start at weight 1. If the song loses the lyrics or the structure, lower the weight. A LoRA trained with a trigger word needs that word in the description.

Common mistakes

Avoid this, do this instead

  • Avoid: Words on the same line as a section tag.

    [verse] Morning light filtering through the pine

    Instead: The tag alone on its line, the lyrics on the lines below it.

    [verse]
Morning light filtering through the pine

  • Avoid: A sung song whose description never names the voice.

    Acoustic pop, 96 BPM, C major, fingerpicked guitar and soft piano.

    Instead: The singer, its timbre, and its delivery, every time.

    Acoustic pop, 96 BPM, C major. Vocals: soft female lead, close and breathy, light stacked harmonies in the chorus. Arrangement: fingerpicked guitar and soft piano; brushed drums and upright bass enter in the chorus.

  • Avoid: Lyrics written into the prompt.

    Instead: The description in the prompt, the words in Lyrics.

  • Avoid: A real artist, band, or song, or "in the style of" someone.

    Instead: The genre, era, instruments, voice, and production that make that sound.

  • Avoid: Lyrics from an existing song.

    Instead: Original lyrics.

  • Avoid: A full song of lyrics with a short Length.

    Instead: Lyrics sized to the Length: a verse and a chorus fill about a minute.

  • Avoid: An open prompt that leaves every decision to the model.

    a happy pop song

    Instead: The genre, tempo, key, mood, singer, and arrangement, decided.

    Upbeat indie pop, 118 BPM, D major. Sunny and carefree, building to a bright, clap-along chorus. Vocals: bright female lead, playful in the verses, belted in the chorus with group "oh-oh" backing vocals. Arrangement: jangly electric guitar and handclaps open; bass and drums enter in the first verse; a glockenspiel hook doubles the chorus melody.

  • Avoid: An @ in the prompt or the lyrics.

    Instead: Words only: Mage reads @ followed by letters as a saved handle, and stops the generation if you have not saved it.

FAQ

MiniMax Music 3 prompting questions

Where do the lyrics go?
In the Lyrics setting on the Advanced page, one sung line per line, with section tags such as [verse] and [chorus] alone on their own lines. The prompt is only the description of the music.
Why did my song come out without vocals?
Either Lyrics was empty, Instrumental was on, or the description never named a singer. Name the voice and its delivery in the prompt, and check that the lyrics are in the Lyrics setting.
Why are some of my lines missing?
Words written on the same line as a section tag are dropped. Put each tag alone on its line, with the sung lines below it. A song can also stop at the Length before the last lines.
How do I make an instrumental?
Turn on Instrumental or leave Lyrics empty, and spend the prompt on the arrangement: the instruments in the order they enter, the key, and how it ends.
Why is my song shorter than the Length I picked?
The Length is a limit, not a target. The model ends the song when the lyrics and the arrangement are done. For a longer song, write more sections: another verse, a bridge, a final chorus.
Can MiniMax Music 3 sing in other languages?
It sings the language of the lyrics. MiniMax demonstrates English and Mandarin Chinese; other languages may work less well, so name the language and the accent in the vocal details and try a short song first.
Can I give it a song or a voice to copy?
No. MiniMax Music 3 takes no audio references, so it cannot copy a voice, a melody, or a recording. Describe the sound and the singer instead.
Does it follow the BPM and key I write?
It follows them closely as guidance, along with the section tags, but not as strict rules: tempo, key, and structure may drift from what you wrote.

Make your first audio with MiniMax Music 3

New accounts get 300 free Gems, enough to try MiniMax Music 3. No install, no credit card.

Try MiniMax Music 3
CharactersCommunity