Ming 3.0 Models by InclusionAI
Last refreshed 2026-10-20. Next refresh: weekly.
Capabilities
About
inclusionAI Ming 3.0 is Ant Group's open-weight speech generation released October 19, 2026 as five models in three lines: Ming-3.0-tts (0.5B dense and Tiny MoE text-to-speech), Ming-3.0-audio (Flash and Tiny audio-language MoE models with optional speech output) and Ming-3.0-asr (Tiny speech recognition). All are MIT-licensed on Hugging Face and ModelScope.
Decision facts
- Best fit
- text to speechaudiospeech recognition
- Capability starting point
- Ming-3.0-audio-flash with multimodal inputs
- Lowest tracked input
- Not tracked
- Closest related family
- Agents-A1
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
Use when the workload needs text to speech, 500M parameters, and audio.
Use when the workload needs text to speech and audio.
Use when the workload needs audio and multimodal inputs.
Use when the workload needs audio and multimodal inputs.
Use when the workload needs speech recognition and audio.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Ming-3.0-tts-0.5B | Use when the workload needs text to speech, 500M parameters, and audio. | 2026-10 | text to speech500M parametersaudio | Current |
| Ming-3.0-tts-tiny | Use when the workload needs text to speech and audio. | 2026-10 | text to speechaudio | Current |
| Ming-3.0-audio-flash | Use when the workload needs audio and multimodal inputs. | 2026-10 | audiomultimodal inputs | Current |
| Ming-3.0-audio-tiny | Use when the workload needs audio and multimodal inputs. | 2026-10 | audiomultimodal inputs | Current |
| Ming-3.0-asr-tiny | Use when the workload needs speech recognition and audio. | 2026-10 | speech recognitionaudio | Current |
Release Timeline
1 release groupSpecifications(5 models)
| Model | Released | Parameters | Multimodal |
|---|---|---|---|
| Ming-3.0-tts-0.5B | 2026-10 | 0.5B | No |
| Ming-3.0-tts-tiny | 2026-10 | 7.9B (1.3B activated) | No |
| Ming-3.0-audio-flash | 2026-10 | 126B (6.5B activated) | Yes |
| Ming-3.0-audio-tiny | 2026-10 | 9.3B (2.7B activated) | Yes |
| Ming-3.0-asr-tiny | 2026-10 | 8.6B (2B activated) | No |