# AI Audio Models: Voice, Music & Sound Effects | Clip Studio Audio Lab > Compare the 8 AI audio models in Clip Studio Audio Lab: ElevenLabs and Gemini text-to-speech, Lyria 3.5 and MiniMax Music 3 for music, and Mirelo SFX 2.0 for sound effects. Canonical page: https://clipstudio.ai/ai-audio-models ## What it is Audio Lab puts narration, music, and sound effect models in one workspace. Pick the job, pick the model, see the quote, and download the audio. There is no single best model: start from the job, then the voice, length, or vocals you need. ## Voice A script read by a voice you choose. - [Eleven Flash v2.5](https://clipstudio.ai/ai-audio-models/eleven-flash-v2-5.md): Fast, natural text-to-speech narration. Maker: ElevenLabs. Voices: 12 named Clip Studio voices. Script: Up to 5,000 characters. Output: Downloadable MP3. - [Eleven v4 Turbo](https://clipstudio.ai/ai-audio-models/eleven-v4-turbo.md): Newer v4 delivery at Flash speed. Maker: ElevenLabs. Voices: 12 named Clip Studio voices. Script: Up to 5,000 characters. Output: Downloadable MP3. - [Eleven v4](https://clipstudio.ai/ai-audio-models/eleven-v4.md): The most expressive ElevenLabs read. Maker: ElevenLabs. Voices: 12 named Clip Studio voices. Script: Up to 5,000 characters. Output: Downloadable MP3. - [Gemini 3.8 Flash TTS](https://clipstudio.ai/ai-audio-models/gemini-3-8-flash-tts.md): Google text-to-speech with 30 prebuilt voices. Maker: Google. Voices: 30 prebuilt Gemini voices. Script: Up to 5,000 characters. Output: Downloadable MP3. - [Gemini 3.8 Flash Lite TTS](https://clipstudio.ai/ai-audio-models/gemini-3-8-flash-lite-tts.md): The lighter Gemini voice model. Maker: Google. Voices: 30 prebuilt Gemini voices. Script: Up to 5,000 characters. Output: Downloadable MP3. ## Music A track from a description, with or without lyrics. - [Lyria 3.5](https://clipstudio.ai/ai-audio-models/lyria-3-5.md): Full songs or instrumentals from a description. Maker: Google DeepMind. Length: Set in your prompt. Vocals: Your lyrics, as described, or instrumental. Output: Downloadable MP3. - [MiniMax Music 3](https://clipstudio.ai/ai-audio-models/minimax-music-3.md): Songs with your own lyrics, up to five minutes. Maker: MiniMax. Length: 30s · 1 min · 2 min · 3 min · 5 min. Vocals: From your lyrics, or instrumental. Output: Downloadable MP3. ## Sound effects Foley and ambience from a description. - [Mirelo SFX 2.0](https://clipstudio.ai/ai-audio-models/mirelo-sfx-2.md): Foley, ambience, and hits from a description. Maker: Mirelo. Length: 3s · 5s · 10s · 20s · 30s. Prompt: Up to 2,000 characters. Output: Downloadable MP3. ## FAQs ### Which AI audio models are in Clip Studio Audio Lab? 8 models. Voice: Eleven Flash v2.5, Eleven v4 Turbo, Eleven v4, Gemini 3.8 Flash TTS, Gemini 3.8 Flash Lite TTS. Music: Lyria 3.5 and MiniMax Music 3. Sound effects: Mirelo SFX 2.0. ### Which voice model should I start with? Eleven Flash v2.5 for a fast, natural read in the 12 named Clip Studio voices. Eleven v4 for the most expressive delivery, Eleven v4 Turbo for v4 at Flash speed, and the Gemini models for Google's 30 prebuilt voices. ### Which music model should I use? Lyria 3.5 for songs or instrumentals described in a prompt. MiniMax Music 3 for a song sung from your own lyrics at an exact length, up to five minutes. ### Are the Audio Lab models free? No. Every plan includes Audio Lab and every model in it. Each run is quoted before you generate and uses your monthly generation allowance. Eligible new accounts get one free Topic Short to try Clip Studio first. ### Do I need separate accounts with ElevenLabs, Google, MiniMax, or Mirelo? No. You use every model from your Clip Studio account, and the audio lands in your Clip Studio library. ## Related - [Audio Lab](https://clipstudio.ai/audio-lab.md) - [AI video models](https://clipstudio.ai/ai-video-models.md) - [AI image models](https://clipstudio.ai/ai-image-models.md)