NVIDIA Nemotron 3.5 Transcribe and Granary v2: Intelligent Multilingual Transcription You Can Prompt and Adapt

NVIDIA Nemotron 3.5 Transcribe is a new open-weight, promptable model that converts recorded speech into useful text. It integrates multilingual transcription across 40 languages, speaker attribution, translation, and audio question answering into a single checkpoint. The model's output can be guided by natural language prompts, and it has achieved top rankings on public leaderboards for accuracy. To facilitate custom adaptation, NVIDIA also released the Granary v2 data pipeline and NeMo Speech fine-tuning tools for domain-specific applications like medical transcription.