The A.I. Daily
NVIDIA releases Nemotron 3.5 Transcribe model and Granary v2 pipeline
NVIDIA released Nemotron 3.5 Transcribe, an open-weight promptable speech model combining transcription, speaker diarization, translation and audio Q&A in one checkpoint. It transcribes 40 languages, translates 31 to or from English and diarizes 13, handling recordings up to an hour with eight speakers. It ranks first among open-weight models on Hugging Face's English ASR leaderboard, alongside the Granary v2 data pipeline and NeMo fine-tuning recipes for domain adaptation.