AI BriefinginclusionAI Releases Ming-3.0-audio-tiny, a 9.3B Parameter Native Audio-Language Model
·2026.10.19 23:29
Key point
The compact Mixture-of-Experts model supports speech recognition, audio reasoning, and optional synthesized speech responses in Chinese and English.
Details
- inclusionAI released Ming-3.0-audio-tiny, a compact native audio-language model designed for efficient audio intelligence.
- The model features 9.3B total parameters with 2.7B activated per inference step, utilizing a Mixture-of-Experts architecture.
- It accepts audio and text inputs to perform speech recognition, spoken question answering, audio understanding, and audio reasoning.
- Capabilities include handling noisy environments, multi-speaker recordings, long-form audio, and non-speech sounds like music and acoustic scenes.
- Outputs are provided as text or text with optional synthesized speech using built-in or reference voices.
- The model supports Chinese and English, including varied accents and dialects, and is released under the MIT license.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
ArtCraft releases open-source IDE for AI image and video creation with 62-model catalog