Voice & vocals
Licensed-data music generation from the leader in AI voice.
ElevenMusic is ElevenLabs' music line, launched in August 2025 with licensing deals from Merlin and Kobalt so its model trains only on licensed data and outputs are cleared for commercial use. It generates tracks from 10 seconds to 10 minutes with or without vocals, and Music v2 (May 2026) added genre transitions inside a track, section inpainting, better vocals and API price cuts of up to 50%. ElevenLabs has since expanded it into a fan-facing ElevenMusic service for creating, remixing and streaming. Music shares a credit pool with ElevenLabs' voice products, and commercial rights start at the cheapest paid plan, except film, TV and games, which need Enterprise. ElevenMusic is the presenting sponsor of the Zinstrel Challenge.
Suno pushed a round of MIDI improvements to Suno Studio on Thursday, announced on the company's Discord. MIDI-to-audio transcription now draws on stronger musical context and stays closer to the source, and audio stems generated from MIDI clips via the chat bar are more accurate.
Not all “AI music” is the same. Some tools write full songs, others just remix, master, or clone voices. Here’s how to tell which ones truly use AI — and which just sound like they do.
While ElevenLabs has gone from voice to song generation with ElevenMusic, Suno's Speech product now takes the platform into ElevenLabs' home turf of voice generation.
ElevenMusic pushed out Music v2.5 today, its new default model for prompted and reference generation, alongside two policy changes that matter more than the model bump itself: a downloads feature and a rights statement.
Universal Music Group and ElevenLabs announced a multi-year licensing agreement and strategic collaboration Wednesday
