NVIDIA has released Nemotron 3.5 Transcribe, an open-weights speech LLM built for transcription. One model handles 40 languages and recordings up to 1 hour. A text instruction picks the output style for each request. Options include verbatim, written form, speaker-attributed, contextual and translated output. For developers, it can replace separate ASR, diarization and alignment tools.
TL;DR
- Size: 34B total and about 4.3B active parameters; 65,536-token context; audio up to 1 hour.
- Runs on: 1x H100 80 GB in BF16. FP8 (about 35 GB) and NVFP4 (about 21 GB) checkpoints exist; NVFP4 runs on an RTX 5090. GGUF files target NeMo-Speech.cpp.
- Performance: Best open-weights model on the English and multilingual Open ASR Leaderboards; top commercial APIs still lead.
- Best: 4.71% mean WER on the 9 public English Open ASR sets (next open model: 4.94%).
- Worst: 34.95% cpCER on AliMeeting Mandarin far-field meetings.
- Bottom line: Best thing: 1 open model covers transcripts, speaker labels, timestamps and translation. Worst thing: offline only, with weak Mandarin meeting accuracy.
What is Nemotron 3.5 Transcribe?
Nemotron 3.5 Transcribe is a speech-augmented language model (SALM) from NVIDIA that turns audio and instructions into text. An audio encoder feeds embeddings into an LLM decoder, which makes transcription promptable. NVIDIA’s technical research report argues there is no single correct transcript. Numbers, fillers and domain terms depend on who reads the output. NVIDIA positions the model as a transcription specialist, not a general-purpose assistant.
How does the architecture work?
It has three parts:
- Parallel Expert Encoder: a trainable 0.6B Transformer ASR encoder runs beside a frozen Nemotron 3 Diarization model. The diarizer tracks up to 8 speakers. Its speaker activity is added to the ASR features every 80 ms. The LLM therefore sees who spoke when, without architecture changes.
- LLM backbone: Nemotron 3.5 Lightning 30B-A3B, a hybrid Mamba-Transformer MoE. It has 52 layers: 23 Mamba-2, 23 MoE and only 6 attention layers. Each MoE layer routes to 6 of 128 experts plus 1 shared expert. Few attention layers keep the KV cache small for hour-long audio.
- Add-ons: a shared multi-token prediction (MTP) layer drafts tokens for speculative decoding. A 50M-parameter CTC adapter aligns the finished transcript to audio for word timestamps.
The model name says 30B-A4B, but the checkpoint holds about 34B parameters. The extra weight comes from the 1.3B MTP module and the audio components.
Training used 1 supervised fine-tuning stage on 256 H100 GPUs for 28,500 steps. A short RL stage followed: 400 GSPO updates with a reward of 1 minus WER. The model saw about 1.72M unique hours of audio, or 6.5M hours with repeats.
How does it score on benchmarks?
All numbers below are self-reported by NVIDIA in the technical report.
- English Open ASR (9 public sets): 4.71% mean WER. The next open model, Qwen3-ASR-1.7B, scores 4.94%. Seven commercial systems still lead, topped by OtterAI speech-v3.1 at 3.88%.
- Multilingual Open ASR (7 languages): 4.62% mean WER, versus 6.49% for Qwen3-ASR-1.7B. ElevenLabs Scribe v2 reaches 3.56%.
- Long-form ASR: 10.98% mean WER, versus 13.37% for Parakeet-TDT-0.6B-v3.
- Multi-speaker, full sessions: 15.80% mean cpWER, versus 16.83% for DiCoW v3.3 with DiariZen. NVIDIA says it is the first SALM to beat cascaded systems here.
- Contextual ASR: an entity list cuts named-entity WER on ContextASR-Bench from 12.21% to 2.20%.
‘}).join(”);BN.textContent=b.n;
setTimeout(function(){BB.querySelectorAll(‘.fl’).forEach(function(f){f.style.width=f.dataset.w+’%’})},40);setTimeout(post,80);}
R.querySelectorAll(‘[data-b]’).forEach(function(c){c.addEventListener(‘click’,function(){R.querySelectorAll(‘[data-b]’).forEach(function(x){x.classList.remove(‘on’)});c.classList.add(‘on’);drawBars(c.dataset.b)})});
drawBars(‘en’);
var P=[
[‘Audio transcription. Language: as spoken.nWording: verbatim, including fillers and repetitions.nFormatting: punctuation and capitalization.nOutput: transcript only.’,’So, uh, the the launch moved tonnext Tuesday at three thirty.’],
[‘Audio transcription. Language: as spoken.nWording: omit fillers and accidental repetitions.nNumbers in their standard written form.’,’So, the launch moved tonnext Tuesday at 3:30.’],
[‘Transcribe with punctuation and capitalization.nContext terms: Nemotron, NVFP4, DFlash2.’,’We tested Nemotron in NVFP4nwith the DFlash2 draft model.’],
[‘Transcribe this audio with punctuation and capitalization.nPut speaker tags <0><1><2><3><0><1>””””””””””””ä1>0>3>2>1>0>
