ModelRadar
/

Ming-3.0-tts-tiny

Ant Group (inclusionAI)๐Ÿ‡จ๐Ÿ‡ณ CNMajor labยท Sep 22, 2026

Audioopen weights

Ming-3.0-tts-tiny 7.9B-A1.3B multilingual MoE speech synthesis with paralinguistic and free-form control.

From the model card

License
mit
Modalities
text โ†’ audio

1 variant

StandardHF

Sources

HuggingFace

Provenance

Detected
Oct 21, 2026
First source
huggingface
Description
From the model card
Editorially reviewed
โ€”