Vocal Synth Models

Explore the evolution and capabilities of ACE Studio's vocal synthesis models.

Model Generation Evolution

ACE Studio continuously iterates its vocal synthesis models. Building on state-of-the-art AI architectures, we incorporate proprietary innovations to enhance naturalness, expressiveness, multilingual accuracy, and voice cloning fidelity.

We also responsibly expand our vocal datasets to cover more languages and musical styles. When architectural improvements and new training data meet our release criteria, we publish a new model generation.

Model List

Below is the current lineup of model generations and variants:

Model Name	Description	Language Support
Verse25	Advanced vocal synth model with high performance and quality.	English, Spanish, Italian, French, Portuguese, Korean, Chinese, Japanese
Verse24	A vocal synth model enhanced with English and Spanish capabilities.	English, Spanish, Chinese, Japanese
Solo24	A vocal synth model expanded from Verse24, designed for voice cloning.	English, Chinese, Japanese
Verse23B	Enhanced based on the Verse23 model.	English, Chinese, Japanese
SoloD	A vocal synth model expanded from Verse23, designed for voice cloning.	English, Chinese, Japanese
Solo23	A vocal synth model expanded from Verse23, designed for voice cloning.	English, Chinese, Japanese
Verse23	ACE Studio's first multilingual vocal synth model.	English, Chinese, Japanese