Skip to main content
Suno music generation models available through Infery. Generate full songs from text prompts with optional custom lyrics, style control and vocal settings.

​
Music models

ModelSlugMax durationFormatsOperations
Suno V4suno-v4240smp3, wavgenerate, extend, upload_cover, upload_extend, vocal_removal, lyrics
Suno V4.5suno-v4-5240smp3, wavgenerate, extend, upload_cover, upload_extend, vocal_removal, lyrics
Suno V4.5 Plussuno-v4-5-plus240smp3, wavgenerate, extend, upload_cover, upload_extend, vocal_removal, lyrics, add_vocals, add_instrumental
Suno V5suno-v5240smp3, wavgenerate, extend, upload_cover, upload_extend, vocal_removal, lyrics, add_vocals, add_instrumental, sounds
Suno V5.5suno-v5-5240smp3, wavgenerate, extend, upload_cover, upload_extend, vocal_removal, lyrics, add_vocals, add_instrumental, sounds

​
Operations

  • generate — Create a new song from a text prompt. Use custom_mode: true for full control over style, title and lyrics.
  • extend — Continue an existing track from a specific timestamp (continue_at). Requires audio_id.
  • upload_cover — Create a cover version from an uploaded audio file (upload_url).
  • upload_extend — Extend an uploaded audio file.
  • add_vocals — Add vocals to an instrumental track (V4.5 Plus+).
  • add_instrumental — Add instrumental backing to a vocal track (V4.5 Plus+).
  • vocal_removal — Separate vocals from a track. Use separation_type: "split_stem" for up to 12 stems.
  • sounds — Generate short sound effects (V5+). Supports sound_loop, sound_tempo, sound_key.
  • lyrics — Generate lyrics from a prompt (low cost).

​
Custom mode parameters

When custom_mode: true:
  • title — Track title (max 100 chars)
  • style — Genre/style description (max 1000 chars)
  • lyrics — Full lyrics text
  • instrumental — Set true for instrumental only
  • negative_tags — Styles to exclude
  • vocal_gender — "m" or "f"
  • persona_id / persona_model — Apply a saved voice persona

​
Tuning parameters

  • style_weight (0–1) — How closely to follow the style prompt
  • audio_weight (0–1) — How much input audio influences the output
  • weirdness_constraint (0–1) — Creativity/novelty level