Nemotron 3.5 Transcribe: #1 open-weight model on Open ASR (English)
Why it matters
NVIDIA released Nemotron 3.5 Transcribe on 28 October 2026, an open-weight, promptable speech model on Hugging Face and ModelScope that combines transcription, speaker attribution, translation and audio question answering in one checkpoint: it transcribes 40 languages, translates 31 of them to or from English, attributes speakers in 13, and handles recordings up to an hour long with up to eight speakers. NVIDIA reports it is first among open-weight models on the Hugging Face Open ASR Leaderboard (English) at 5.33% average WER, and ties for first among open-weight models on Artificial Analysis at 2.8% AA-WER with about 4B active parameters versus 24B for the model it ties. Caveat: it is built for recorded audio, not live streaming.
Read the source: huggingface.co (opens in a new tab)
All hits from 30 Oct 2026More VoiceGet the daily hits by email