Nvidia Nemotron 3.5 Transcribe: open ASR for 40 languages

Nvidia has released Nemotron 3.5 Transcribe, an open-weight AI model that performs transcription, speaker attribution, translation, and audio Q&A across 40 languages, with top performance on multiple leaderboards. The model can be customized using natural language prompts and fine-tuned for domain-specific tasks, with Nvidia also releasing Granary v2, an open data pipeline and fine-tuning recipes to help developers adapt the model to their specific needs. Quantized versions will enable local deployment on consumer and edge devices, while the full model scales to data center GPUs.