AI News / Audio

inclusionAIOpen weightsPrimary source

Ming-3.0-tts arrives as an MIT-licensed speech family

A 0.5B dense model and a tiny 7.9B-total Mixture-of-Experts model with voice cloning and voice design across nine languages and dialects. Exact day not verified.

Why it matters: It gives builders a highly ranked open option. The license is MIT, though the README adds a usage disclaimer against impersonation.

Parameters
0.5B dense; 7.9B total / 1.3B active
Weights size
3.0 GB for the 0.5b repo
License
MIT
BenchmarkScoreTypeMeasured or published by
AA Pronunciation Robustness93.2%, 1st overallCitedREADME, citing Artificial Analysis
AA Controlled Voice1069 Elo, 1st among openCitedREADME, citing Artificial Analysis
Sources: github.com

Written by AI. Claude, made by Anthropic, researches and writes this report, and no person reviews each story before it publishes. Anthropic is one of the companies covered. Check the linked sources for anything important. How this works.