Skip to content

MiniMax Music 2.6

MiniMax's song model: give it a style and lyrics and it composes a complete song of the length it judges right, in about a minute and a half.

Updated November 9, 2026

A humpback whale gliding through golden clouds above a 1950s seaside town at dusk

Example image. Write what you want to see and press Create. The image appears in the same frame within seconds. Video, music and voice work the same way.

At a glance

At a glance
MakerMiniMax
Strengths in the appFull songs
Who can use itPaid plans
How it’s paidCredits, from the first use
PrivacyPrivate: an open model run by NeverNo on providers chosen because they don’t keep requests. Your history stays encrypted in your browser.

Price

What the app charges, read from the same table the servers bill with. 1 credit = $0.01.

Price
Generation38 credits ($0.38)

What it is

MiniMax Music 2.6 is a music model from MiniMax, available in the Music tab of NeverNo's Audio Studio. The Studio tags it song. Where the ACE-Step models make a track of a length you choose, this one decides for itself: you describe the style, add lyrics of up to 3,000 characters if you wish, and it composes a full song with a beginning, a middle and an end.

A full song takes about 90 seconds to make, as measured on Runware, our audio provider, against about 6 seconds for ACE-Step 1.5 Turbo. That is the trade. You wait longer and have no length setting, but you are asking a model built around the idea of a finished song rather than a clip, which suits lyrics that have verses, a chorus and a closing line.

Be clear about the rules. MiniMax Music 2.6 is a moderated model: it follows its maker's rules, and requests that MiniMax does not accept are declined. It is not one of the private models in the Studio, so for music served by NeverNo's own provider with no outside lab involved, use ACE-Step 1.5 or ACE-Step 1.5 Turbo. NeverNo's own hard limits apply to every model. The finished song is stored in your browser, not on our servers.

Audio is paid in credits by any signed-in account, so a paid plan is not required. A practical routine is to sketch the genre and the lyric on ACE-Step 1.5 Turbo, then bring the same text here when you want the model to shape it into a complete song. Plan for the wait and for a single decisive attempt. Since a render takes around a minute and a half and the length is out of your hands, it pays to prepare the input: polish the lyric, mark the sections, state the genre, the voice and the feel in one clear line, and avoid contradictory instructions such as a slow ballad with fast dance energy. When you listen back, judge the whole arc, the opening, the lift into the chorus and the final line, and revise the text rather than the style if something feels rushed or dragged. Typical results fit stories that need a beginning and an end: a lullaby, a football chant, a farewell song for a colleague, a theme for a tabletop campaign. Share the finished file with friends, because it is yours to download from your browser, and keep the lyric text alongside it so that you can rerun a better version later. Think of the result as a demo you can actually play at a gathering, a birthday party or a launch, rather than a loop for background use, and expect natural variation each time you run the same lyric: melodies, tempo and instrumentation can all differ, so generating twice and choosing is perfectly reasonable. Teachers, couples planning a ceremony and small clubs often want a personal song rather than a catalog track, and a named place, a nickname or an inside joke in the lyrics makes the result memorable. Keep the lyric sheet and your favorite takes in a folder, so favorite versions never vanish behind newer experiments. Listen with headphones before sharing, checking vocals, rhythm and the closing bars.

Strengths

  1. Composes a complete song and chooses its length itself.

  2. Lyrics up to 3,000 characters, with a style description.

  3. Tagged song, the model of the Studio built around a finished track with verses and a chorus.

  4. Open to any signed-in account, paid in credits, with no paid plan needed.

Limits

  1. About 90 seconds to make a song, against about 6 for ACE-Step 1.5 Turbo.

  2. No length setting: the model decides how long the song is.

  3. Moderated by its maker: requests that MiniMax does not accept are declined, and it is neither anonymized nor private.

When to choose it

  • A finished song from your lyrics

    Write the words of a birthday song, a wedding toast or a band's demo, name a style, and receive a complete track with a proper ending rather than a clip that stops.

  • Songs for a brand or a channel

    A short jingle-like song for a bakery, a podcast theme with sung words, or a signature track for a streamer, composed to your wording.

  • Finding the structure of a lyric

    Hand it a long poem and hear where the model puts the chorus and the pauses; then rewrite the lines that did not scan.

Prompt tips

  1. State the genre, the voice and the energy in one line, then paste lyrics with section labels such as verse, chorus and bridge.

  2. Keep each lyric line short and singable; the model composes around them.

  3. Because the model chooses the length, a longer lyric gives a longer song, and a short one a shorter song. Repeating a chorus in your text signals the hook you want the melody to build around, while a one-line bridge adds contrast without stretching the runtime.

Frequently asked questions

How long does MiniMax Music 2.6 take?

About 90 seconds for a full song, measured on Runware, our audio provider.

Can I choose the length?

No. The model chooses the length of the song.

Can I write my own lyrics?

Yes, up to 3,000 characters.

Is MiniMax Music 2.6 private?

No. It is a moderated model that follows MiniMax's rules; for music with no outside lab involved, use ACE-Step 1.5 or ACE-Step 1.5 Turbo. Your tracks are stored in your browser.

Do I need a paid plan?

No. Audio is paid in credits by any signed-in account.

Try MiniMax Music 2.6 in the Studio

Open the Studio