Suno 6

Suno 6 (suno-music-generate, short names suno and suno-v6; the old suno-v5 keeps working) generates vocal songs from a text prompt. It returns 2 tracks per request. Generation is async, like video: create a job, poll the status, download the finished mp3 files. The model is paid — requests are charged from your balance.

Parameters

A job is created with POST /v1/audio/music:

Field	Description
model	suno or suno-v6 (the full key is suno-music-generate; suno-v5 is a legacy synonym)
prompt	Track description, up to 5000 characters; required in simple mode
mode	simple (from prompt) or custom (from lyrics + style + title); defaults to simple
lyrics	Song lyrics; required in custom mode, up to 5000 characters
style	Genre, style, tags; up to 500 characters
title	Track title; up to 200 characters

A request with mode: simple and no prompt, or mode: custom and no lyrics, returns 400.

Examples

Simple mode

Python

import time

import requests

API = "https://api.mixen.ai/v1/audio/music"
headers = {"Authorization": f"Bearer {MIXEN_API_KEY}"}

job = requests.post(
    API,
    headers=headers,
    json={"model": "suno", "prompt": "Melancholic synth-pop about a night city, female vocals"},
).json()

while job["status"] in ("queued", "in_progress"):
    time.sleep(5)
    job = requests.get(f"{API}/{job['id']}", headers=headers).json()

for i in (0, 1):  # Suno yields two tracks for a single charge
    mp3 = requests.get(f"{API}/{job['id']}/content", params={"index": i}, headers=headers)
    with open(f"track_{i}.mp3", "wb") as f:
        f.write(mp3.content)

cURL

curl https://api.mixen.ai/v1/audio/music \
  -H "Authorization: Bearer $MIXEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "suno",
    "prompt": "Melancholic synth-pop about a night city, female vocals"
  }'

Custom mode

When you need specific lyrics rather than generated ones:

{
  "model": "suno",
  "mode": "custom",
  "title": "Night Tram",
  "style": "indie rock, 90s, male vocals, live drums",
  "lyrics": "First verse...\n\nChorus..."
}

Poll and download

The statuses are the same as for video: queued → in_progress → completed or failed; a finished answer carries a tracks array with index and title.

curl https://api.mixen.ai/v1/audio/music/music_... \
  -H "Authorization: Bearer $MIXEN_API_KEY"

curl -L "https://api.mixen.ai/v1/audio/music/music_.../content?index=1" \
  -H "Authorization: Bearer $MIXEN_API_KEY" -o track_1.mp3

Two tracks per request is how the model behaves, not an error: both appear in the poll response, and a specific track is downloaded with the ?index=N parameter of /audio/music/{id}/content (numbering starts at zero, default 0). The endpoint returns mp3 and supports Range requests. A job that has not finished within 15 minutes is marked failed on the next poll and is not charged.

Price

8.0216 ₽ ($0.09) per request — the per_request_rub field in the catalog. Music is billed per request rather than per second: Suno’s two tracks are one charge.

Strengths and limits

• A vocal song from a text prompt; each request yields two track variants.
• Custom mode: your own lyrics, genre and title.
• Two tracks for a single charge — pick the better one from the poll response.
• Instrumental tracks without vocals are the Lyria model — see the music guide.
• Balance only; a job that times out is not charged.

All models — in the catalog. The full music workflow — music guide.