Nvidia releases Nemotron 3.5 Transcribe, open-weight speech model

What happened: Nvidia has released Nemotron 3.5 Transcribe, a 30-billion-parameter open-weight speech model. It is available on Hugging Face. TechCrunch reported the release on 28 October.

The details: The model transcribes audio in up to 40 languages. It translates 31 languages to and from English. It handles recordings up to one hour long and distinguishes up to eight speakers. It works on recorded audio only, not livestreams.

The hardware: The model can run on a single chip. Supported hardware ranges from Jetson AGX Thor and DGX Spark up to Blackwell datacenter chips.

The rivals: Nvidia claims the model beats rivals on benchmarks, including Hugging Face's Open ASR Leaderboard. The article gives no specific scores.

The context: Nvidia also released version 2 of Granary, its open-source speech dataset. The company's earlier Parakeet model is a multilingual open-weight voice model with 1.1 billion parameters.

The catch: The article calls the model "free, open-weight" but names no specific licence.

What's next: No further release plans were reported.