whatmodel.sh / stt / Nemotron 3.5 ASR Streaming 0.6B
Nemotron 3.5 ASR Streaming 0.6B
↗ share
+ contribute
Nemotron 3.5 ASR Streaming 0.6B is an open-weights speech-to-text model from NVIDIA, released May 2026. Its published size is 0.64B parameters. Licence: openmdw-1.1.
Catalogue identity notes
Exact checkpoint nvidia/nemotron-3.5-asr-streaming-0.6b (NeMo library per HF; streaming-capable multilingual ASR covering 32 locales per the transcribe.cpp catalog; licence OpenMDW-1.1 per that catalog, HF 'other'). Params: HF safetensors total 637,997,088. released: HF repo creation month (2026-05-15). No local command added (not verified here).
Score within the Open ASR English short-form WER (March 2026) comparison set only.
$ how it compares
The speech-to-text models ranked either side of Nemotron 3.5 ASR Streaming 0.6B, out of 176 with a score, then more from NVIDIA.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 139 |  Whisper small.en | 44±3 | local | ||
| 140 |  ESPnet Branchformer 18-layer (70 epochs, no LM) | 44±4 | local | ||
| 141 | Nemotron 3.5 ASR Streaming 0.6B | 44±3 | local | ||
| 142 | Whisper medium | 43±2 | local | ||
| 143 |  Omnilingual ASR CTC 300M v2 | 43±3 | local | ||
| more from NVIDIA | |||||
| 28 | Canary-Qwen-2.5B | 61±2 | local | ||
| 35 | Parakeet TDT 0.6B v2 | 60±1 | local | ||
| 37 | Parakeet Unified EN 0.6B | 60±4 | local | ||
see all speech-to-text models · find the ones that fit your machine
$ what it's good at
other tests8 not in a task · 98% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| LibriSpeech test-clean WER +1 | 3.0% ✎ | ▁···▁▁▂·▄▄██ beats 11/56 | 31% |
| FLEURS English test WER | 7.9% ✎ | ▁··▁▁▁▁▃▁▃█▁ beats 8/56 | 11% |
| Picovoice STT benchmark · English streaming 4-set mean WER | 7.8% ✎ | █·█···▄▄▄··▄ beats 7/7 | 11% |
| FLEURS Mandarin test error rate | 18.9% ✎ | ▂▂··▂▄▂▂·▆██ beats 4/17 | 11% |
| GigaSpeech test WER | 11.8% ✎ | ▁····▁··▃▄█▂ beats 2/17 | 10% |
| LibriSpeech test-other WER | 6.8% ✎ | ▁▁▁▃▅▃▅▄█▂▂▁ beats 6/54 | 9.5% |
| Danish ASR Leaderboard · mean WER over 5 Danish test sets (RyeAI) | 47.4% ✎ | ▂▂···▃▂▅█▃▂▂ beats 2/16 | 8.4% |
| Open Universal Arabic ASR Leaderboard · 6-test-set average WER | 35.9% ✎ | ▃·▂▂▅█▅▃▂▅▂▂ beats 14/22 | 6.3% |
$ where to run it
Other ways to run it
version
- Run it as published · needs ~0.74 GB · from huggingface.co/nvidia (official)
curl -fsSL https://github.com/NVIDIA/NeMo-Speech.cpp/raw/main/scripts/install.sh | sh && export PATH="$HOME/.local/bin:$PATH" && hf download nvidia/nemotron-3.5-asr-streaming-0.6b nemotron-3.5-asr-streaming-0.6b.q8_0.gguf --local-dir models && nemo-speech transcribe audio.wav --model models/nemotron-3.5-asr-streaming-0.6b.q8_0.gguf --language en-US
Official weights and documentation: huggingface.co/nvidia ↗
$ how the score was worked out · transcribe.cpp model catalog… 50%, VibeVoice-ASR-Streaming Techn… 22%, Picovoice speech-to-text-benc… 11% and 2 more
Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works
| source | results | share of the score |
|---|---|---|
 transcribe.cpp model catalog · LibriSpeech test-clean WER of validated full-precision ports | 3 | 50% |
 VibeVoice-ASR-Streaming Technical Report (arXiv:2609.02812v2), Table 3 single-speaker recognition | 3 | 22% |
| Picovoice speech-to-text-benchmark · English batch-engine WER (README) | 1 | 11% |
 Danish Speech-to-Text Leaderboard (RyeAI) · results dataset | 1 | 8% |
| Open Universal Arabic ASR Leaderboard (Elm Research Center) | 1 | 6% |
| prior · pull toward the middle | 2% | |
| weighted average | = 43.5 |
$ improve this page
Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.
data updated · sources and licences · json