aiqbench analysis
This is an open-weight, 30B-parameter mixture-of-experts model with 4B active parameters, aimed at combining speech detection, transcription, speaker labeling, and Q&A in one pass. The claimed speed and accuracy figures, including the 30.8% out-of-box and 11.8% fine-tuned word error rates, come from Nvidia and need independent benchmarking. The practical draw for developers is that the model can run from edge devices to data-center GPUs, and the open fine-tuning pipeline could matter more than the leaderboard result.
Nvidia releases all-in-one adaptable Nemotron 3.5 audio transcription and translation model
Nvidia Corp. today released a leaderboard-winning open artificial intelligence model for multilingual audio transcription and translation, along with an open data pipeline for fine-tuning Nemotron 3.5 Transcribe. Traditional transcription AI models follow a workflow that detects speech, transcribes it, identifies and labels different speakers, then answers questions about the transcription. Transcribe does all of this […] The post Nvidia releases all-in-one adaptable Nemotron 3.5 audio transcription and translation model appeared first on SiliconANGLE.
Nvidia releases all-in-one adaptable Nemotron 3.5 audio transcription and translation model Nvidia Corp. today released a leaderboard-winning open artificial intelligence model for multilingual audio transcription and translation, along with an open data pipeline for fine-tuning Nemotron 3.5 Transcribe. Traditional transcription AI models follow a workflow that detects speech, transcribes it, identifies and labels different speakers, then answers questions about the transcription. Transcribe does all of this in a…
Excerpt shown under the publisher's RSS terms · 68 words · full text and figures are on the publisher's site
Related coverage
Transcribe, Translate, and Optimize: Joint Reward Learning for Speech Translation
arXiv · Sep 23, 2026
HK-LegiCoST: Leveraging Non-Verbatim Transcripts for Speech Translation
arXiv · Jun 17, 2026
Granite Speech 5.0: Transcribe, Translate, and Attribute more Accurately and Faster
Hugging Face · Oct 21, 2026
Learning When to Commit from Partial Speech for End-to-End Simultaneous Speech Translation
arXiv · Oct 5, 2026