Skip to main content
POST
Eleven Music v2.5
eleven-music-v2.5 composes a song or an instrumental from a text prompt, with optional lyrics, and returns the finished audio file in the response. You choose the length, from 3 seconds to 5 minutes, and pay for exactly that length. Confirm live rates on the SnapGen models page. Failed requests are not charged.
string
required
Set to eleven-music-v2.5.
string
required
What the music should sound like, up to 4,100 characters: genre, mood, instruments, tempo, and structure.
string
Lyrics to sing, up to 4,000 characters.
integer
default:"30"
Length in whole seconds, from 3 to 300. Billed as requested.
boolean
true guarantees an instrumental with no vocals.
string
default:"mp3"
mp3, opus, or pcm. See Output formats.

Request

The request waits until the track is finished and returns the audio as the body, so give your client a timeout of several minutes.

Write the prompt

  • Describe the sound: genre, mood, instruments, tempo, and vocal style, for example Uplifting cinematic trailer score with strings and big drums.
  • Add lyrics in lyrics when you want specific words sung, up to 4,000 characters.
  • Force an instrumental with instrumental: true, for background music or beds under speech.
  • Pick the length with duration_seconds. You pay for the length you ask for, so request only what you need.

Price examples

To edit a song in later steps, such as extending, covering, remixing, or splitting stems, generate it with suno-music or flow-music instead. Those models run as tasks that later operations can reference.
Synchronous audio endpoints reject Idempotency-Key. If a request times out, the track may already have been generated: check your Console request logs before you retry. See Music composition for the response and errors.