# AI Music Models: Mureka 9.5, Lyria 3.5, ACE-Step 1.5, MiniMax Music 3 | Clip Studio > Compare the four AI song models in Clip Studio Music Lab: Mureka 9.5, Lyria 3.5, ACE-Step 1.5 and MiniMax Music 3. Vocals, lyrics, length and tempo lock side by side, with real samples. Canonical page: https://clipstudio.ai/ai-music-models ## What it is Music Lab runs four song models in one studio. Pick a model for the song you want, then open the result as a multitrack project: every AI edit lands as a new take, and mixing and mastering are free. ## Song models - [Mureka 9.5](https://clipstudio.ai/ai-music-models/mureka-9-5.md): Full songs with vocals that you can keep editing, line by line. Maker: Mureka. Makes: Lead vocal, or instrumental. Length: 2:00 – 5:00. Lyrics: Yours, or written for you. Tempo & key lock: Set by the model. Edits after: Stems, rewrite a part, extend, re-sing a line, change the style. - [Lyria 3.5](https://clipstudio.ai/ai-music-models/lyria-3-5.md): Rich, full arrangements from a single sentence. No lyrics needed. Maker: Google DeepMind. Makes: Vocals or instrumental. Length: Set by the model, up to 3:30. Lyrics: Written by the model. Tempo & key lock: Set by the model. Edits after: Stems, then edits on the split tracks. - [ACE-Step 1.5](https://clipstudio.ai/ai-music-models/ace-step-1-5.md): Locks the tempo, key and time signature you set. Up to ten minutes. Maker: ACE Studio. Makes: Vocals in 50+ languages, or instrumental. Length: 0:10 – 10:00. Lyrics: Yours, or written for you. Tempo & key lock: Tempo, key and time signature. Edits after: Stems, then edits on the split tracks. - [MiniMax Music 3](https://clipstudio.ai/ai-music-models/minimax-music-3.md): Your own lyrics, sung word for word. Up to five minutes. Maker: MiniMax. Makes: Vocals from your lyrics, or instrumental. Length: 0:30 – 5:00. Lyrics: Yours, or written for you. Tempo & key lock: Set by the model. Edits after: Stems, then edits on the split tracks. ## FAQs ### Which AI music model should I use? Start with Mureka 9.5 for a song with vocals you will keep editing. Pick Lyria 3.5 for a mood from one sentence with no lyrics to write, ACE-Step 1.5 when the tempo, key and time signature must be exact, and MiniMax Music 3 to sing your own lyrics word for word. ### Can every model make instrumentals? Yes. Pick Instrumental as the singer and the song is recorded with no vocals. ### Can I split any song into stems? Yes. Split into stems works on every song in Music Lab, whichever model wrote it. ### Can I use the songs in my videos? Yes. Use your songs in your videos and posts. Exports have no watermark. ### How are the models priced? Every song and every AI edit is quoted before it runs and uses your monthly generation allowance. Editing, mixing and mastering are free. ## Related - [Music Lab](https://clipstudio.ai/music-lab.md) - [AI audio models](https://clipstudio.ai/ai-audio-models.md) - [AI video models](https://clipstudio.ai/ai-video-models.md) - [AI image models](https://clipstudio.ai/ai-image-models.md)