NVIDIA · MODEL RELEASE TRACKER
Nemotron 3.5 Transcribe
Nemotron 3.5 Transcribe is an open-weight, promptable speech model from NVIDIA that performs transcription, translation, speaker attribution/diarization, and audio question answering. It supports many languages, handles recordings up to an hour, and is available on Hugging Face and ModelScope.CURRENT SNAPSHOT3/5 DIMENSIONS WITH DATA
The dimensions that change the decision.
Pricing / licensingDescribed as a free, open-weight model.
Maximum recording lengthHandles recordings up to one hour long.
Modality and featuresSpeech/audio model that is promptable and combines transcription, translation, speaker attribution/diarization, and audio question answering in a single checkpoint.
Not established from the available sources.
Not established from the available sources.
RELEASE TIMELINE
Published, source-backed release events only.
VERIFIABLE FACTS
Every value stays attached to a source and date.
RELEASE · Release and availabilityDEVELOPER CLAIM
Released by NVIDIA as an open-weight model; made available on Hugging Face and ModelScope.
PRICING · Pricing / licensingDEVELOPER CLAIM
Described as a free, open-weight model.
MODALITIES · Modality and featuresDEVELOPER CLAIM
Speech/audio model that is promptable and combines transcription, translation, speaker attribution/diarization, and audio question answering in a single checkpoint.
CAPABILITIES · Transcription language coverageDEVELOPER CLAIM
Transcribes audio in 40 languages.
CAPABILITIES · Translation capabilityDEVELOPER CLAIM
Can translate 31 languages to and from English.
CAPABILITIES · Speaker attribution coverageDEVELOPER CLAIM
Performs speaker attribution (diarization) in 13 languages.
CAPABILITIES · Number of speakers labeledDEVELOPER CLAIM
Labels up to eight speakers in a recording.
CONTEXT WINDOW · Maximum recording lengthDEVELOPER CLAIM
Handles recordings up to one hour long.
WHAT CHANGED