Back to models
M3
Mac and Windows

inclusionAI / Ming-3.0-tts-0.5B

Ming 3.0 TTS

Expressive speech and voice cloning in seven languages

Mac and Windows

Parameters

0.5B

Languages

7

Output

44.1 kHz

Platform

Mac and Windows

Model overview

Built into OpenVox for private local voice generation.

Use in OpenVox

Ming 3.0 TTS 0.5B is a multilingual local speech model with reference voice cloning, voice design, and expressive delivery controls.

The 0.5B checkpoint supports Chinese, English, German, Spanish, French, Korean, and Japanese, plus Sichuanese and Cantonese dialect controls.

Its upstream release documents zero-shot voice cloning, natural-language voice design, and controls for emotion, speed, pitch, and volume. Check the installed OpenVox version for exposed controls.

In the published Seed-TTS comparison, the 0.5B checkpoint reports 0.800% Mandarin error and 1.584% English WER. These are upstream results, not OpenVox benchmarks.

Ming 3.0 TTS 0.5B is available in OpenVox for Mac and Windows after model download.

OpenVox does not provide celebrity voice models and does not permit cloning, impersonating, or commercially using any person's voice without proper rights or consent.

Open-source model: inclusionAI / Ming-3.0-tts-0.5B

View on Hugging Face