Ant Ling Open-Sources the Ming-3.0 Speech Series, Five Models in Total

Artificial Analysis

Ant Group's Ming-3.0-tts achieves top-tier speech synthesis quality with 0.5B parameters

Ant Group has released Ming-3.0-tts, which achieves top-tier text-to-speech quality with 0.5B parameters and is released under the MIT license. Artificial Analysis says it is the only model among the top three on its speech synthesis leaderboard with fewer than 1B parameters; among the top ten of the Controlled Voice Arena, the other open-weight TTS models are all 3.5B or larger, and Ming-3.0-tts reaches the same tier with 7 to 70 times fewer parameters.

Shao Meng

Blogger plans to run Ming-3.0-tts locally as a reading and dubbing assistant

The blogger plans to run the 0.5B Ming-3.0-tts voice synthesis and cloning model locally, to make a book-reading assistant for his son and German dubbing for videos. The reading assistant supports reading through the whole text and giving prompts at pauses or in specified areas; the dubbing solution combines ASR and multimodal video understanding to generate German dubbing and synchronize it with the visuals. He plans to put it into daily use in a week and open-source it.

inclusionai_new_reposFirst-partySelected

English translation is being prepared.

Ant Group's inclusionAI: New GitHub repositoryFirst-party

Ant Group's inclusionAI open-sources the Ming-3.0-audio-flash native audio model

Ant Group's inclusionAI has open-sourced the Ming-3.0-audio-flash native audio model on GitHub. It uses a 126B-parameter MoE architecture with 6.5B activated parameters, and there is also a Ming-3.0-audio-tiny version with 9.3B parameters and 2.7B activated parameters. The models cover Chinese and English recognition across multiple dialects and accents, long audio understanding, audio reasoning, and end-to-end speech dialogue. They support three inference methods: vLLM-Omni, SGLang, and Transformers, and are released under the MIT license.

ithome_ai

English translation is being prepared.

Ant LingSelected

Ant Ling's Ming-3.0 Speech Models Get Day-0 Support in vLLM-Omni

Ant Ling has open-sourced 5 speech models covering TTS, audio understanding, and ASR, with total parameter counts ranging from 0.5B to 126B (6.5B active parameters). vLLM announced that vLLM-Omni provided Ming-3.0 support on the day of release; Ant Ling thanked them for this and said deployment is now more convenient.

AntLingAGI

AntLing Ming-3.0 Speech Models Integrated into atomic.chat Local Workflow

AntLingAGI thanks atomic.chat for integrating Ming-3.0 into local AI workflows, making open-source speech models more accessible. According to atomic.chat, Ming-3.0 comprises 5 open-source models covering speaking, listening, and speech transcription; its speech model ranks No. 1 among open-source models on Artificial Analysis, and the smallest version is only 0.5B.

Artificial Analysis

Ant Group's Ming-3.0-tts Tops the Open-Source TTS Leaderboard

The Ming-3.0-tts model from Ant Group's LingAGI team has become the top-ranked open-weight model in the Artificial Analysis Controlled Speech TTS Arena, and with a pronunciation robustness score of 93.2% it is tied for first among 31 models, level with Fish Audio Drama 3 Preview and ahead of Eleven v4's 91.7%. The model has 0.5B dense parameters, supports preset voices and stateless voice cloning, and its weights are released on Hugging Face under the MIT license; there is also a 7.9B MoE variant, Ming-3.0-tts-tiny. Its Controlled Speech Elo is 1,069, ranking first among open-source models and 13th overall; its Provider Voice Elo is 1,188, ranking third among open-source models and 19th overall.

Ant Ling

Ant Ling Open-Sources the Ming-3.0 Series of 5 Speech Models

Ant Ling has open-sourced 5 models in the Ming-3.0 series, covering three speech families: tts, audio, and asr. Ming-3.0-tts includes a 0.5B speech design and paralinguistic cue model and a Tiny 7.9B A1.3B long-speech model; Ming-3.0-audio includes a Flash 126B A6.5B understanding and tool model and a Tiny 9.3B A2.7B efficient dialogue model; Ming-3.0-asr is Tiny 8.6B A2B, aimed at dialects and context.