Nemotron 3.5 ASR Streaming 0.6B
Nemotron 3.5 ASR Streaming 0.6B is a 600M-parameter speech-to-text model from NVIDIA, released under the OpenMDW-1.1 license. The reviewed source lists 40 language-locales in tiers (28 explicitly enumerated in our record), and the model is designed for native streaming. Among the common setups we test, it ranks highest for live captions and voice input (#1 of 26 in the matcher).
Native low-latency streaming with tiered locale support; some locales are adaptation-ready rather than out-of-the-box transcription-ready.
Key facts
| Provider | NVIDIA |
|---|---|
| Model family | Nemotron |
| Parameters | 600M |
| License | OpenMDW-1.1 · Custom terms — license text ↗ |
| Language coverage | 40 language-locales in tiers · 28 enumerated in our record · 8 adaptation-ready |
| Streaming | Native streaming architecture |
| Timestamps | Limited or runtime-dependent |
| Speaker diarization | No — needs a separate component |
| Speech translation | No |
| Documented runtimes | NVIDIA NeMo, Hugging Face Transformers ≥ 5.13, NeMo-Speech.cpp — how to run it |
| Primary source | huggingface.co/nvidia/nemotron-3.5-asr-streaming-0.6b ↗ |
| Last reviewed |
Hardware and workload fit
These are Speech Model Match's conservative 1–5 editorial ratings, relative to the other 19 reviewed models. They come from model size, official memory guidance and documented runtimes — not from a universal benchmark. How the ratings work.
| Dimension | Rating | Scale |
|---|---|---|
| CPU-only fit | 2/5 | |
| Apple Silicon fit | 3/5 | |
| NVIDIA GPU fit | 5/5 | |
| Speed | 5/5 | |
| Accuracy (relative) | 5/5 | |
| Low memory use | 3/5 | |
| Long recordings | 4/5 | |
| Live / low latency | 5/5 | |
| Subtitle work | 4/5 | |
| Edge / low-power devices | 2/5 |
Where it is strong
- Native streaming architecture for live, low-latency transcription
- NVIDIA GPU fit: rated 5/5
- Speed: rated 5/5
- Accuracy (relative): rated 5/5
- Long recordings: rated 4/5
- Live / low latency: rated 5/5
- Subtitle work: rated 4/5
What to check first
- CPU-only fit: rated 2/5
- Edge / low-power devices: rated 2/5
- No native speaker diarization — pair it with a separate diarization pipeline if you need speaker labels
- Timestamp support is limited or runtime-dependent — test it if you need subtitles
- Custom model license; read the current terms before production or redistribution.
How it ranks in common setups
Ranks come from the same deterministic matcher that runs on the homepage, applied to all 26 reviewed models. They are decision guidance, not accuracy claims.
| Setup | Rank | Fit score | Top match |
|---|---|---|---|
| Long recordings on an 8–12 GB NVIDIA GPU, balanced | #7 of 26 | 75/100 · Strong fit | Whisper turbo |
| Live captions or voice input (native streaming + timestamps, speed first) | #1 of 26 | 91/100 · Excellent fit | This model |
| Subtitles with timing, accuracy first | #10 of 26 | 78/100 · Strong fit | Whisper medium |
| Meetings with speaker labels (diarization + timestamps) | #11 of 26 | 75/100 · Strong fit | VibeVoice-ASR 7B |
| CPU-only machine, low memory use | #16 of 26 | 58/100 · Possible fit | VibeVoice-ASR-BitNet |
| Apple Silicon with 8–16 GB unified memory, long recordings | #15 of 26 | 66/100 · Possible fit | Whisper small |
| Japanese live transcription (native streaming, speed first) | #1 of 26 | 91/100 · Excellent fit | This model |
Supported languages
NVIDIA groups language-locales into tiers. Languages marked as adaptation-ready need fine-tuning before full transcription.
- Arabic
ar - Bulgarian
bg - Croatian
hr - Czech
cs - Danish
da - Dutch
nl - English
en - Estonian
et - Finnish
fi - French
fr - German
de - Hindi
hi - Hungarian
hu - Italian
it - Japanese
ja - Korean
ko - Mandarin Chinese
zh - Norwegian Bokmål
nb - Polish
pl - Portuguese
pt - Romanian
ro - Russian
ru - Slovak
sk - Spanish
es - Swedish
sv - Turkish
tr - Ukrainian
uk - Vietnamese
vi
Adaptation-ready (fine-tuning required)
Compare coverage across every reviewed model in the language support matrix.
Community activity on Hugging Face
- Downloads, last 30 days
- 539,841
- Likes
- 1,269
- Repository last updated
- 10 Sept 2026
- Published on the Hub
- 15 May 2026
Figures from the Hugging Face Hub API, 11 Nov 2026. Source: huggingface.co/nvidia/nemotron-3.5-asr-streaming-0.6b ↗. Downloads count file downloads, not active users, and do not affect the matcher's scores. Compare all models.
How to run Nemotron 3.5 ASR Streaming 0.6B locally
NVIDIA documents NeMo as the reference runtime and native Transformers support for this streaming model. Runtimes below are the ones the official model card or repository documents for this model (checked November 2, 2026).
- NVIDIA NeMo ↗Framework · reference runtime engine
- Hugging Face Transformers ≥ 5.13 ↗Framework · native RNN-T support
- NeMo-Speech.cpp ↗Official package · C++ runtime, q8_0 GGUF
Install command shown in the official huggingface.co page ↗:
pip install "transformers>=5.13.0"
Audio input: Mono WAV input. Streaming chunk sizes of 80, 160, 320, 560 or 1120 ms are documented; maximum file length depends on GPU memory.
Official hardware guidance: Listed hardware: NVIDIA Ampere, Blackwell, Hopper, Jetson, Lovelace, Turing and Volta on Linux. No CPU-only or Apple Silicon guidance is published. Source ↗
Whatever runtime you choose, measure speed and accuracy on your own audio with the benchmark tool; results differ a lot between runtimes, precision and hardware.
Before you deploy Nemotron 3.5 ASR Streaming 0.6B
A printable checklist built from the facts on this page. Tick it off before putting the model into production.
- Language. Confirm your language is in the 28-language list on this page, and test any accents or code-switching you expect.
- License. OpenMDW-1.1: When you redistribute, include a copy of the license and keep copyright and origin notices. Rights end if you sue claiming the model materials infringe a patent or copyright. Places no restrictions on outputs. License guide.
- Install. Start from the official command:
pip install "transformers>=5.13.0"(full run notes). - Audio input. Mono WAV input. Streaming chunk sizes of 80, 160, 320, 560 or 1120 ms are documented; maximum file length depends on GPU memory.
- Hardware. Listed hardware: NVIDIA Ampere, Blackwell, Hopper, Jetson, Lovelace, Turing and Volta on Linux. No CPU-only or Apple Silicon guidance is published.
- Features you rely on. For Nemotron 3.5 ASR Streaming 0.6B, native streaming is documented; timestamps are limited or runtime-dependent; speaker labels need a separate diarization component. Test each one you need with real audio.
- Measure. Run 10–20 representative clips, score them with the WER calculator and record speed in the benchmark workspace before you commit.
Alternatives and related models
Similar fit profile: Qwen3-ASR 0.6B (Qwen, 0.6B), Parakeet TDT 0.6B v3 (NVIDIA, 600M), Voxtral Mini 4B Realtime 2602 (Mistral AI, 4B). Similarity is computed from the ten fit ratings plus streaming and diarization support.
| Fact | Nemotron 3.5 ASR Streaming 0.6B | Qwen3-ASR 0.6B | Parakeet TDT 0.6B v3 |
|---|---|---|---|
| Provider · size | NVIDIA · 600M | Qwen · 0.6B | NVIDIA · 600M |
| License | OpenMDW-1.1 Custom terms | Apache-2.0 Permissive | CC BY 4.0 Attribution required |
| Languages | 40 language-locales in tiers | 30 languages + 22 Chinese dialects | 25 European languages |
| Streaming | Native | Via runtime | Via runtime |
| Timestamps | Limited | Separate aligner | Yes |
| Speaker diarization | Separate pipeline | Separate pipeline | Separate pipeline |
| CPU · Apple · NVIDIA fit | 2/5 · 3/5 · 5/5 | 2/5 · 3/5 · 5/5 | 2/5 · 3/5 · 5/5 |
| Live · long recordings | 5/5 · 4/5 | 4/5 · 4/5 | 2/5 · 5/5 |
| Main runtime | NVIDIA NeMo | qwen-asr package | NVIDIA NeMo |
Head-to-head pages: Nemotron 3.5 ASR Streaming 0.6B vs Voxtral Mini 4B Realtime 2602
Compare Nemotron 3.5 ASR Streaming 0.6B side by side
Guides that cover Nemotron 3.5 ASR Streaming 0.6B
- Parakeet v3 vs Nemotron 3.5 ASR Streaming
- Best local STT models for an 8 GB GPU
- Best local Whisper alternatives in 2026
- Best local speech-to-text models in 2026
- Best local speech-to-text models for multilingual use
- Best local speech-to-text models for streaming
Check Nemotron 3.5 ASR Streaming 0.6B for your setup
Pick your language, hardware and workload to see this model's fit score, the reasons behind it and where it ranks against all reviewed models. Everything runs in your browser.
Common questions
Can I use Nemotron 3.5 ASR Streaming 0.6B commercially?
Custom model license; read the current terms before production or redistribution. The listed license is OpenMDW-1.1; read the current license text ↗ before shipping a product, because terms can change. The license guide compares the obligations of all 26 reviewed models.
Does Nemotron 3.5 ASR Streaming 0.6B support real-time streaming?
Yes. The reviewed source describes a native streaming architecture, so it is a candidate for live captions, dictation or voice agents. Measure end-to-end latency on your own hardware.
Which languages does Nemotron 3.5 ASR Streaming 0.6B transcribe?
40 language-locales in tiers. The full list of 28 explicitly enumerated languages is shown above, plus 8 adaptation-ready languages that need fine-tuning.
Can Nemotron 3.5 ASR Streaming 0.6B tell speakers apart?
Not natively. Combine it with a dedicated diarization component and evaluate that component separately; the meetings guide covers this workflow.
Primary sources
Reviewed against primary sources on September 16, 2026. Model behavior, runtime support, language coverage and licenses can change; re-check the linked primary source before production use. See the full source directory.