IT之家 · ITHome
Ming-3.0 goes open source with five speech and audio models
The open-source Ming-3.0 family comprises five models for speech synthesis, audio understanding and dialogue, and speech recognition. Developers can select models for individual tasks or combine them in workflows. Its synthesis models support voice design using text or reference audio and accept performance directions. Audio models can transcribe, summarize and analyze viewpoints in long recordings, returning structured results.
What to know
- 01
Ming-3.0-tts includes a director mode for specifying emotional shifts, word emphasis and pacing.
- 02
Ling Studio offers APIs for Ming-3.0-tts-0.5B, Ming-3.0-audio-flash and Ming-3.0-asr-tiny.
Why it matters
The family offers models that developers can combine to connect speech input with spoken responses in a workflow.
AI-assisted brief. Read the source for full details. Editorial & sourcing policy
