GoKawiilNvidia releases open-weight Nemotron 3.5 Transcribe speech model
Nvidia released Nemotron 3.5 Transcribe, a free, open-weight transcription model supporting up to 40 languages, alongside version 2 of its open-source Granary speech dataset. The company says the 30-billion-parameter model can translate 31 languages to and from English, process recordings up to an hour long, distinguish up to eight speakers, and run on a single chip ranging from Jetson AGX Thor to Blackwell datacenter hardware. Nvidia claims it outperforms rivals on benchmarks including Hugging Face's Open ASR Leaderboard, though it only works on recorded audio rather than live streams.
GoKawiil's interpretation of the reporting above, not reported fact.
The release targets business meetings and medical transcription, sectors where AI has already gained workplace adoption, suggesting Nvidia is positioning its open models as infrastructure for developers rather than competing directly in consumer voice apps. It arrives amid a wave of voice AI launches from Mistral, ElevenLabs and Salesforce, indicating intensifying competition to set standards in speech transcription and translation. Offering the model and dataset openly could help Nvidia encourage developers to build on its hardware ecosystem across edge and datacenter chips.
- Nemotron 3.5 Transcribe supports 40 languages and translates 31 to/from English
- The model runs on a single chip, from Jetson AGX Thor to Blackwell datacenters
- Release follows recent voice AI launches from Mistral, ElevenLabs and Salesforce
Source: techcrunch.com — Julie Bort, 2026-10-28
Published there as: “Voice AI in vogue: Nvidia releases new open transcription model”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.