None
None
Long Story Video Skill
Turn your idea into 10s–10min video—AI writes scripts, prompts, and generates footage automatically.
Ads Video Skill
Generate professional ads and sales videos—AI auto-generates scripts, prompts, and footage.
3D Science Explainer Video Skill
Convert scientific concepts into stunning 3D explain animations
Feedback
freeTrialImage.bannerPity
freeTrialImage.upgradeUnlock
- ✓freeTrialImage.benefitHd
- ✓freeTrialImage.benefitWatermark
- ✓freeTrialImage.benefitUnlimited
nemotron transcribe
Nemotron Transcribe turns audio into instruction-guided transcripts — 40 languages, one-hour recordings, speaker labels and word timestamps.
All Tools
Discover our comprehensive AI-powered animation toolkit
Muse Video
Seedance2.5
The Future of AI Video Is Here.
Seedance 2.0
The Future of AI Video Is Here.
Veo3.1
Create Stunning Videos with Veo3.1
Grok Video Generator
Create Videos from Text or Images with AI
MiniMax H3 video generator
MiniMax H3 AI Video Generator
Seedance 2.0
The Future of AI Video Is Here.
Gemini Omni 1.2 Flash
What Nemotron Transcribe Does Differently
No single transcript format suits every reader. Nemotron Transcribe lets you describe the convention you need in plain language — and then follows it.
- Transcription Styles Set by InstructionVerbatim, written-form, contextual or speaker-attributed output — the style is chosen per request, simply by what the instruction asks for.
- Speech-Augmented Language Model DesignPairs a speech encoder and a frozen diarization model with a 30B-A3B hybrid MoE LLM running at roughly 4.3B active parameters.
- Open Weights Plus a Fine-Tuning ToolkitOpenMDW-1.1 weights arrive in BF16, FP8, NVFP4, GGUF and DFlash2 formats, together with the NeMo Speech toolkit for Nemotron Transcribe.
Running Nemotron Transcribe in Three Steps
Serve a checkpoint, write the instruction, then decode and align — that is the entire path to an attributed transcript.
Core Capabilities of Nemotron Transcribe
Leaderboard-topping accuracy, speaker attribution for overlapping talk and fast NIM inference — the strengths Nemotron Transcribe delivers.
Leaderboard-Topping ASR Accuracy
A mean WER of 4.71% across the nine English Open ASR test sets, and 4.62% on the multilingual board.
Attribution That Handles Overlapping Speakers
From one-minute clips to full meetings, it is the first SALM to outperform cascaded systems on attributed ASR.
Contextual Transcription Driven by Entity Lists
Supplying expected entities cuts named-entity errors six-fold on ContextASR-Bench, powered by the parallel expert encoder.
Per-Speaker Word Timestamps
A compact CTC adapter maps the finished transcript back onto the audio, so long overlapping calls still return timings for every speaker.
High-Throughput, Efficient Inference
The model's MTP layer or a DFlash2 draft model speeds up decoding; the NVFP4 NIM processes 2,238 audio seconds per second on a single RTX PRO 6000 Blackwell.
New Tasks Adapted on a Single GPU
A supervision-balanced mixture lets one GPU absorb new languages and domains in roughly 0.5–2.3 GPU-hours per task.
Nemotron Transcribe: Frequently Asked Questions
Straight answers about Nemotron Transcribe — benchmarks, language coverage, timestamps, licensing and the checkpoints you can download.
What exactly is Nemotron 3.5 Transcribe?
It is a speech-augmented language model that combines a speech encoder, a frozen diarization model and a 30B-A3B hybrid MoE LLM in one system.
Which transcription styles can I ask for?
Verbatim, written-form, contextual, speaker-attributed and rich transcription, plus translation and language ID — each one selected through an instruction.
How accurate is it on public benchmarks?
It scores 4.71% mean WER on the nine English leaderboard sets and 4.62% multilingual, while making six times fewer named-entity errors on ContextASR-Bench.
How long can a recording be, and how many languages are covered?
A single Nemotron Transcribe checkpoint handles recordings of up to one hour across 40 languages, code-switched speech included.
Can it show who spoke and when?
Yes. Overlapping multi-speaker audio is attributed per speaker, and per-speaker word timestamps come back whenever you request them.
What license and checkpoints ship, and can I fine-tune?
OpenMDW-1.1 weights are released as BF16, FP8, NVFP4, GGUF and DFlash2 files, and NeMo Speech fine-tunes them in about 0.5–2.3 GPU-hours per task.
Put Nemotron Transcribe to Work Today
Pick a checkpoint, serve it through the Nemotron Transcribe NIM, and send your first instruction-driven request — styles, translation, language ID and speaker timestamps across 40 languages.