MiniMax Music 3 vs Lyria 3.5
MiniMax Music 3 (by MiniMax) and Lyria 3.5 (by Google DeepMind) are both voice and music models that run side by side in AI Whisperer Lab. MiniMax Music 3: 5 min songs, 44.1 kHz stereo WAV, 9 section tags, 20 cr per minute. Lyria 3.5: 3 min songs, 44.1 kHz stereo, SynthID watermark, 16 cr per song.
Side by side
| MiniMax Music 3 | Lyria 3.5 | |
|---|---|---|
| Made by | MiniMax | Google DeepMind |
| Its numbers | 5 min songs | 3 min songs |
| 44.1 kHz stereo WAV | 44.1 kHz stereo | |
| 9 section tags | SynthID watermark | |
| 20 cr per minute | 16 cr per song |
What MiniMax Music 3 nails
- Up to 5 minutes — Complete songs up to 300 seconds; the default is 60 seconds.
- Shape every section — Tags like [verse], [chorus], [bridge] and [outro], each on its own line, set the arrangement.
- Uncompressed audio — 44.1 kHz, 16-bit stereo WAV.
What Lyria 3.5 nails
- Vocals that sound real — Google calls Lyria 3.5's vocals realistic, natural and nuanced, and they can be in different languages.
- Your lyrics or its own — Write lyrics with [Verse] and [Chorus] tags, let it write them, or ask for an instrumental.
- Studio quality — 44.1 kHz stereo, with an inaudible SynthID watermark in every track.
Questions people ask
Is MiniMax Music 3 or Lyria 3.5 better?
It depends on the job. Pick MiniMax Music 3 for up to 5 minutes; shape every section; uncompressed audio. Pick Lyria 3.5 for vocals that sound real; your lyrics or its own; studio quality. Both run in the same Lab studio, so you can run one prompt on each and keep the better result.
Can I use MiniMax Music 3 and Lyria 3.5 without two subscriptions?
Yes. In AI Whisperer Lab both run on one credit balance with no subscription: credits never expire, the exact price shows before every run, and a failed run gives its credits back.