inclusionAI releases five open-weight Ming-3.0 audio models for speech and transcription

inclusionAI releases Ming-3.0, five open-weight audio models covering controllable speech, long-recording analysis and context-aware transcription, according to ModelScope's announcement on X. ModelScope says Ming-3.0-tts-0.5B reaches a 1069 Elo rating on Artificial Analysis Controlled Voice, the top open-weight result, using eight English reference voices.

ModelScope also reports that Ming-3.0-audio-flash scores 99.40 on BigBenchAudio S2T in the authors' own comparison, and that Ming-3.0-asr-tiny accepts hotwords and dialogue history.