AI music generators compared: prompts, limits and structure

Last checked 4 October 2026

Seven AI music generators, seven ways of reading a prompt. This page sets them side by side, first by what each platform documents, then by how APMonster writes for each one. Then it writes one brief for all seven, so you can see what changes and why.

What does each platform document?

Each cell is a fact the platform's own pages give, as its guide sets out with the source and the date it was read.

Platform	Prompt length	Things to avoid	Sung lyrics	Longest song	Downloads
Suno	None published (its form took 1,000 characters)	Exclude Styles box	Yes (its form took 5,000 characters)	8 minutes per generation	Yes, capped by plan
Eleven Music	4,100 characters (API)	Exclude styles, per track or section	Yes	10 minutes	Yes
Lyria 3.5	131,072 tokens (API)	None documented	Yes, under a Lyrics: header	About 3 minutes	MP3 or WAV (API)
Stable Audio 3	10,000 characters (API)	None on the hosted API	No: instrumentals only	380 seconds (API)	Yes, and you own them
EasyMusic	1,000 characters (Style)	Exclude styles box	Yes (5,000 characters)	Set from 10 to 360 seconds (Fusion)	MP3, M4A, WAV and stems
Loudly	1,000 characters	None	MANTA 1 only (3,000 characters)	420 seconds (free plan: 30)	WAV, MP3 and stems
Udio	None stated	Style Reduction	Yes	2:10 per generation, extendable	Off since 29 October 2025

How does APMonster write for each platform?

Each row is built from the file APMonster writes every prompt from. APMonster's working length is how long APMonster keeps its prompts for that platform, never more than the platform's own limit where it states one.

Platform	Prompt form	APMonster's working length	Things to avoid	Song structure	LyrAssist writes	Tempo
Suno	Short sentences	900 characters	"Exclude Styles"	Tags in the lyrics	Sung lyrics	"N BPM"
Eleven Music	Section plan	1,500 characters	"Exclude styles"	Section plan with lengths	Sung lyrics	"N BPM"
Lyria 3.5	A topic or genre, then comma-separated tags	1,200 characters	None: said positively	Timestamped segments	Sung lyrics	"N BPM"
Stable Audio 3	Track type, then comma-separated fields	1,500 characters	None: said positively	Described in words	A vocal guide for a singer	"N BPM"
EasyMusic	Short sentences	900 characters	"Exclude Styles"	Tags with cues, in the lyrics	Sung lyrics	"N BPM"
Loudly	Short sentences	1,000 characters	None: said positively	Described in words	A vocal guide for a singer; sung lyrics for MANTA 1	"N BPM"
Udio	A topic or genre, then comma-separated tags	1,500 characters	"Style Reduction"	Tags in the lyrics	Sung lyrics	"N BPM"

One brief, written seven ways

Here is one song, briefed once. Below it is the prompt APMonster's format calls for on each platform, and each structure as APMonster writes it. Each guide explains what changed for its platform, and why.

The brief

Genre: dream pop
Era: late 1980s
Tempo: 92 BPM
Key: A minor
Mood: wistful, hopeful
Instruments: chorus-soaked electric guitar, warm analog synth pads, soft drum machine, round bass
Vocals: breathy female lead
Avoid: heavy distortion, trap hi-hats
Structure: Intro (4 bars, instrumental), Verse (16 bars), Chorus (8 bars), Verse (8 bars), Chorus (8 bars), Bridge (8 bars), Chorus (8 bars), Outro (4 bars, instrumental); about 2:47

Suno

Style

Dream pop at 92 BPM in A minor. A breathy female lead, wistful and hopeful. Chorus-soaked electric guitar, warm analog synth pads, a soft drum machine and round bass, with a late-1980s shimmer.

Exclude Styles

heavy distortion, trap hi-hats

Lyrics (the section tags; LyrAssist writes the lines under each)

[Intro]
[Verse 1]
[Chorus]
[Verse 2]
[Chorus]
[Bridge]
[Chorus]
[Outro]

Eleven Music

Overall styles (the start of the prompt)

Global styles: dream pop, late-1980s shimmer, 92 BPM, A minor, wistful, hopeful, chorus-soaked electric guitar, warm analog synth pads, soft drum machine, round bass, breathy female vocal

Exclude styles

heavy distortion, trap hi-hats

Sections (the rest of the prompt; LyrAssist adds the lyrics)

[Intro] (10s) style: soft, instrumental
[Verse 1] (42s) style: steady
[Chorus] (21s) style: strong
[Verse 2] (21s) style: steady
[Chorus] (21s) style: strong
[Bridge] (21s) style: soft
[Chorus] (21s) style: full force
[Outro] (10s) style: very soft, instrumental

Lyria 3.5

Prompt

dream pop, chorus-soaked clean electric guitar, warm analog synth pads, soft steady drum machine, round bass, breathy female vocals, 92 BPM, A minor, wistful, clean warm mix

Timed segments (the rest of the prompt)

[0:00 - 0:10] Intro: instrumental. Intensity: 3/10 (Low)
[0:10 - 0:52] Verse 1. Intensity: 5/10 (Medium)
[0:52 - 1:13] Chorus. Intensity: 8/10 (High)
[1:13 - 1:34] Verse 2. Intensity: 5/10 (Medium)
[1:34 - 1:55] Chorus. Intensity: 8/10 (High)
[1:55 - 2:16] Bridge. Intensity: 4/10 (Low)
[2:16 - 2:37] Chorus. Intensity: 9/10 (Very high)
[2:37 - 2:47] Outro: instrumental. Intensity: 2/10 (Very low)

Stable Audio 3

Prompt

TrackType: Music, VocalType: Instrumental, Genre: dream pop, Instruments: chorus-soaked electric guitar, warm analog synth pads, soft drum machine, round bass, Moods: wistful, hopeful, slow build into each chorus, clean, warm late-1980s production, 92 BPM

The arc (added to the prompt)

Opens with a soft instrumental intro, builds into a steady verse, surges into a strong chorus, drops to a steady verse, surges into a strong chorus, drops to a soft bridge, surges into a full-force chorus, and ends on a very soft instrumental outro.

Duration setting

2:47

EasyMusic

Style

Late-1980s dream pop at 92 BPM, wistful but hopeful. A breathy female voice leads over chorus-soaked electric guitar and warm analog synth pads, with a soft drum machine and round bass underneath. It opens sparse, swells into each chorus and fades out.

Exclude styles

heavy distortion, trap hi-hats

Lyrics (the section tags and cues; LyrAssist writes the lines under each)

[Instrumental Intro — soft]
[Verse 1 — steady]
[Chorus — strong]
[Verse 2 — steady]
[Chorus — strong]
[Bridge — soft]
[Chorus — full force]
[Instrumental Outro — very soft]
[End]

Target duration

2:47

Loudly

Prompt

Wistful dream pop at 92 BPM with a late-1980s shimmer and a gently building energy. Chorus-soaked guitar and warm analog synth pads float over a soft drum machine and round bass in a clean, airy mix.

The arc (added to the prompt)

Opens with a soft instrumental intro, builds into a steady verse, surges into a strong chorus, drops to a steady verse, surges into a strong chorus, drops to a soft bridge, surges into a full-force chorus, and ends on a very soft instrumental outro.

Length setting

2:47

Udio

Prompt

A slow drive home through a sleeping city, dream pop, breathy female vocals, wistful, hopeful, chorus-soaked electric guitar, warm analog synth pads, soft drum machine, round bass, 92 BPM, late-1980s shimmer

Style Reduction

heavy distortion, trap hi-hats

Lyrics (the section tags; LyrAssist writes the lines under each)

[Intro]
[Instrumental]
[Verse]
[Chorus]
[Verse]
[Chorus]
[Bridge]
[Chorus]
[Outro]
[Instrumental]

The structure blocks are exactly what the app writes for this song. The prompts show the shape APMonster asks for on each platform; the wording of a generated prompt varies from one generation to the next.

Frequently asked questions

Which AI music generators take a negative prompt?

Suno has an Exclude Styles box and EasyMusic an Exclude styles box, Udio has Style Reduction, and Eleven Music has Exclude styles for the whole track or one section. Google documents no negative prompt for Lyria 3.5, Stability's hosted API has none for Stable Audio 3, and Loudly has none: for those, describe what you want instead, as APMonster does.

Which AI music generators make instrumentals only?

Stable Audio 3 makes instrumentals only. Loudly's VEGA models do too, while its MANTA 1 model sings.

Where do section tags like [Chorus] go?

On Suno, EasyMusic and Udio they go in the lyrics box. Eleven Music takes a plan with a length for each section, Lyria takes section tags or timestamps in the prompt, and Stable Audio 3 and Loudly take no section tags, so the song's arc is described in words.

Which AI music generator makes the longest songs?

Eleven Music, at up to 10 minutes, then Suno at up to 8 minutes in one generation. Loudly makes up to 420 seconds, Stable Audio 3 up to 380 seconds through its API, and Lyria 3.5 about 3 minutes. Udio makes 2:10 at a time and extends from there.

Can you download songs from every platform?

No. Udio turned off downloads on 29 October 2025, and Suno limits them by plan. Eleven Music, Stable Audio 3, EasyMusic and Loudly offer downloads, and Lyria 3.5 returns MP3 or WAV through its API.