AI News/Wire summary
Nvidia releases open Nemotron 3.5 Transcribe model for 40 languages
Key points
- Nemotron 3.5 Transcribe is free, open-weight, and transcribes audio in up to 40 languages, released Wednesday.
- Nvidia also released version 2 of Granary, its open source speech dataset.
- Nvidia claims the model translates 31 languages to and from English and handles recordings up to an hour.
- Transcribe works only with recorded audio, not livestreams; Nvidia says it has 30 billion parameters and runs on one chip.
Why it matters for marketers: Teams building meeting or medical transcription apps gain a free option to tune for industry jargon, though it cannot handle live audio.
Summary by the Z Media news desk, from the outlets' reporting. Read the full stories at the source.
Read the original
TechCrunch: Voice AI in vogue: Nvidia releases new open transcription model
More coverage
- Hugging Face: NVIDIA Nemotron 3.5 Transcribe and Granary v2: Intelligent Multilingual Transcription You Can Prompt and Adapt
- The Decoder: Nvidia's open Nemotron 3.5 Transcribe combines transcription, translation, and speaker recognition
- MarkTechPost: NVIDIA Releases Nemotron 3.5 Transcribe: An Open 30B MoE Model for 40-Language Transcription with ~4B Active Parameters
Related on Z Media
- Google ships Gemini 4 Carbon and Genie 4, both in limited accessOur brief, via Google DeepMind, 5 days ago
- Anthropic's IPO filing shows a fast-growing ad budget and a bet on enterprise buyersOur brief, via The Verge, 6 days ago
- Mistral Large 4 released on Hugging Face with open weightsvia Hugging Face, 4 days ago
- Seekr eBook proposes a new human oversight framework for AI agentsvia Google DeepMind, 5 days ago
- OpenAI's GPT-6.1 Terra cuts API prices in half and hallucinates lessOur brief, via OpenAI, Oct 25
- Google's Suncatcher space data center prototype is online and communicatingvia TechCrunch, 2 days ago