NVIDIA Releases Open-Weight Nemotron 3.5 Transcribe Model Supporting 40 Language Locales and Granary v2 Dataset
What happened
NVIDIA released Nemotron 3.5 Transcribe, a free open-weight speech model supporting 40 language locales alongside an expanded 463,062-hour open dataset.
The Context
- Multi-task architecture consolidates your stack: The release handles transcription, translation across 31 languages to and from English, and speaker diarization for up to eight speakers in recordings reaching one hour.
- Granary v2 sets a new benchmark for open audio training: NVIDIA paired the model release with 463,062 hours of open-source multilingual audio across 48 languages to accelerate developer fine-tuning.
- Enterprise-ready processing reduces pipeline latency: Built on a cache-aware FastConformer-RNNT architecture, the tool processes both real-time streaming and batch workloads with native punctuation and capitalization.
The Number
— The number of language locales supported out of the box, allowing your product to scale localized voice capabilities across international markets from day one.
Takeaway
Expect developer adoption of open-weight audio models to accelerate as teams replace paid transcription APIs with self-hosted, multi-task alternatives.
Daily AI news: models, tools, and policy.
"Independent outlet tracking the fast pace of AI."
— A47 Editor
Nvidia's open Nemotron 3.5 Transcribe combines transcription, translation, and speaker recognition
Nvidia has released Nemotron 3.5 Transcribe, an open speech recognition model that handles transcription, speaker recognition, and translation and takes its instructions as text prompts. According to The Decoder, the model leads several benchmarks wh...
Curated tech headlines including AI stories.
"Influential aggregator surfacing the day’s top tech/AI links."
— A47 Editor
Nvidia debuts Nemotron 3.5 Transcribe, an open-weight model for multilingual audio transcription and translation, and Granary v2, its open-source speech dataset (Kyt Dotson/SiliconANGLE)
Nvidia Corp. has released Nemotron 3.5 Transcribe, an open-weight model for multilingual audio transcription and translation, alongside Granary v2, its open-source speech dataset, according to a report by Kyt Dotson of SiliconANGLE. The company says ...
Research, news, and analysis on blockchain startups, DeFi, and regulations.
"Crypto Briefing provides research, news, and analysis on blockchain startups, DeFi, and crypto regulations with investor-focused coverage."
— A47 Editor
Nvidia releases Nemotron 3.5 ASR, a free open-weight model for 40-language audio transcription
Nvidia has released Nemotron 3.5 ASR, a free open-weight model that supports audio transcription across 40 languages, according to Crypto Briefing. The model is made available openly rather than as a paid, closed service.
AI news with an enterprise and cloud focus.
"Covers AI in the context of data infrastructure, cloud, and enterprise stacks."
— A47 Editor
Nvidia releases all-in-one adaptable Nemotron 3.5 audio transcription and translation model
Nvidia Corp. has released Nemotron 3.5 Transcribe, an open, all-in-one audio transcription and translation model that the company says topped a leaderboard, along with an open data pipeline for fine-tuning it. Unlike traditional workflows that separa...
Startup news with frequent AI coverage.
"Covers launches, funding, and product updates in AI."
— A47 Editor
Voice AI in vogue: Nvidia releases new open transcription model
Nvidia has released a free, open speech-to-text model capable of transcribing up to 40 languages, according to TechCrunch. The model works on recorded audio rather than live audio, positioning it for meeting and medical transcription tools while leav...
Open-source AI tools, research, and ecosystem updates.
"Core hub for transformers, datasets, and open models."
— A47 Editor
NVIDIA Nemotron 3.5 Transcribe and Granary v2: Intelligent Multilingual Transcription You Can Prompt and Adapt
NVIDIA has published details of its Nemotron 3.5 Transcribe and Granary v2 models, which the Hugging Face blog describes as multilingual transcription systems that users can prompt and adapt. The announcement, categorized under AI, carries a neutral ...