NVIDIA released Nemotron 3.5 Transcribe, an open-weight promptable speech model for 40-language transcription, diarization, translation and audio Q&A, plus the

So what

What this event means by reading role—not a longer recap.

BuilderYou can self-host one checkpoint for transcription, speaker labels, translation and audio Q&A, and steer output with natural-language prompts instead of separat
ResearcherIt pairs a hybrid Mamba-Transformer MoE (~4B active of 30B params) with Sortformer speaker cues, and Granary v2 opens 463k hours across 48 languages for adaptat
InvestorAn open-weight model leading open ASR and diarization benchmarks at ~4B active parameters pressures proprietary speech API pricing and distribution.

Also covering this story

2 sources clustered into one event.

Read original Back to Stories

Score dimensions

Higher total means read first. Each bar is one factor we use to rank the system pool. How we score

RelevanceHow tightly this is about AI.: 88
ImpactHow much this could change the field or the market.: 80
NoveltyHow new this is versus a recap.: 82
CredibilityHow much we trust the source.: 100
ActionabilityWhether a reader can do something with it.: 85