All tasks / Models
Nemotron 3.5 ASR Streaming 0.6B
NVIDIA's small open-weights streaming speech recognition model for low-latency voice agents.
- NVIDIA
- Open weights
- Released 15 May 2026
Where it ranks
- CTranscriptionModels · Local · 9.8% on Hugging Face Open ASR Leaderboard (English, short-form average WER) (secondary benchmark)
Facts
- Vendor
- NVIDIA
- Family
- Nemotron ASR Streaming
- Released
- 15 May 2026
- Weights
- Open: download and run it yourself
- License
- OpenMDW-1.1
- Parameters
- 0.638B
- GPU memory to run locally
- ~2 GB (Derived: about 0.64B parameters is roughly 1.3 GB of bf16 weights plus runtime overhead; not an official requirement.)
- Access
- Free (open weights)
- Last verified
- 19 Oct 2026
Benchmark results
| Benchmark | Result | Entry on the leaderboard | As of | Source |
|---|---|---|---|---|
| Coval Benchmarks: Speech-to-Text WER (24h window) | 12.9% | together / nemotron-3.5-asr-streaming-0.6bCoval STT benchmark, mean WER over the trailing 24h window (snapshot generation 2610, as_of 2026-10-19T09:38Z), n=899 clips; rank 31 of 31; read from the public API that feeds https://benchmarks.coval.ai. | 19 Oct 2026 | benchmarks-api-6wxgp27p2a-ue.a.run.app |
| Hugging Face Open ASR Leaderboard (English, short-form average WER) | 9.8% | nvidia/nemotron-3.5-asr-streaming-0.6bAverage WER over the board's default datasets (English short-form, version 17-10-2026); rank 51 of 56; read from the Space's /config table data. | 17 Oct 2026 | huggingface.co |
Where you can use it
- BNVIDIA NeMo SpeechEngines & APIs · Local
Earlier versions
- nemotron-speech-streaming-en-0.6b · —