NVIDIA Releases Nemotron 3.5 Transcribe with Instruction-Driven ASR
NVIDIA released Nemotron 3.5 Transcribe, a speech LLM built on a hybrid Mamba-Transformer MoE architecture. It integrates a frozen diarizer with an ASR encoder to enable instruction-driven transcription, allowing users to control output formats (verbatim, written, translation) via text prompts. While it outperforms Qwen3-ASR in English benchmarks, it is offline-only and shows lower accuracy for Mandarin far-field meetings.
Full Report
▲ Marktechpost AI · 2026-10-29 04
▲ NVIDIA just released Nemotron 3.5 Transcribe, a single speech LLM that writes transcripts the way you ask for them.
The key idea is instruction-driven transcription. A frozen diarizer feeds "who spoke when" into the ASR encoder, and a hybrid Mamba-Transformer MoE LLM writes the output. One text instruction picks verbatim, written form, context terms, speaker tags or translation, with word timestamps on request.
On the 9 public English Open ASR sets it scores 4.71% mean WER, ahead of Qwen3-ASR-1.7B at 4.94%. Multilingual is 4.62% versus 6.49%. It is a trade-off though: it is offline only, commercial APIs still lead (OtterAI at 3.88%), and Mandarin far-field meetings sit at 34.95% cpCER.
34B total and about 4.3B active parameters, 40 languages, audio up to 1 hour. BF16 runs on 1 H100; NVFP4 fits an RTX 5090. OpenMDW-1.1 license....
Full analysis:
