NVIDIA Releases Nemotron 3.5 Transcribe Open-Weight Speech Model, Supporting 40 Languages
On October 28, NVIDIA released Nemotron 3.5 Transcribe, an open-weight speech model that can transcribe audio in 40 languages, translate speech, and identify speakers, and developers can control the transcription format with natural language instructions without retraining. The model consists of a Transformer speech encoder, a speaker diarization model, and the Nemotron 3.5 Lightning language model, with approximately 4.3 billion activated parameters, and the released checkpoint is labeled 30B-A4B. NVIDIA says it achieves an average word error rate of 4.71% on nine public test sets of the English Open ASR Leaderboard and 4.62% on seven languages of the multilingual leaderboard, and calls it the best open-weight model on both leaderboards; the model is released in BF16, FP8, and NVFP4 formats as well as GGUF files, under the OpenMDW-1.1 license.
Why read this
AI score 68
The original text presents capabilities such as multilingual transcription, speaker diarization, and word-level timestamps, and lists WER and inference throughput data to help assess its applicable scenarios.
Topics
Body
NVIDIA released Nemotron 3.5 Transcribe on October 28, an open-weight speech model that can transcribe audio in 40 languages, translate speech and identify speakers. Developers can use natural-language instructions to control transcript formatting without retraining the model, according to NVIDIA.
The model supports verbatim transcripts, cleaned-up written text, domain-specific terminology and code-switched speech. It can process recordings of up to one hour and generate speaker-attributed word-level timestamps on request. NVIDIA notes that individual capabilities vary by language and task.
Nemotron 3.5 Transcribe combines a Transformer speech encoder and a speaker diarisation model with NVIDIA’s Nemotron 3.5 Lightning language model. The architecture has approximately 4.3 billion active parameters. Its released checkpoints use the designation 30B-A4B.
NVIDIA reported a mean word error rate (WER) of 4.71% across nine public test sets on the English Open ASR Leaderboard. Its mean WER on the multilingual Open ASR Leaderboard was 4.62%, covering seven languages. A lower WER indicates fewer transcription errors. NVIDIA describes the model as the best open-weight model on both leaderboards.
The company also reported a six-fold reduction in named-entity errors on ContextASR-Bench when the model received a list of expected entities. This feature is intended to improve the transcription of names and specialised terms.
For inference speed, NVIDIA reported throughput of up to 2,238 seconds of audio per second on a single RTX PRO 6000 Blackwell GPU. This result uses the NVFP4 checkpoint, DFlash2 speculative decoding and NVIDIA’s Nemotron 3.5 Transcribe inference software.
However, these serving optimisations are not yet part of the open-weight release. NVIDIA says the release-day vLLM implementation is expected to run at roughly half the speed of the optimised setup. Actual performance will depend on the hardware, configuration and workload, according to the company.
NVIDIA has released model checkpoints in BF16, FP8 and NVFP4 formats, along with GGUF files for NeMo-Speech.cpp. The model is available under the OpenMDW-1.1 licence, and a fine-tuning toolkit is available through NeMo Speech.
The company says developers can adapt the model to new languages, domains and tasks. Its technical report describes experiments across 17 tasks, with adaptation typically taking between 0.5 and 2.3 GPU-hours per task.