NVIDIA releases promptable Nemotron 3.5 Transcribe and Granary v2 speech data
NVIDIA’s open-weight Nemotron 3.5 Transcribe combines transcription, speaker attribution, translation and audio question answering. Granary v2’s public manifest spans 463,062 hours across 48 languages and 16 source corpora. Developers can serve the checkpoint with vLLM, but it supports recorded audio rather than live streaming.
Why it mattersBuilders can use one checkpoint for recorded meeting transcripts, speaker labels and translations, then adapt recognition to domain vocabulary with NeMo Speech fine-tuning tools.