Model Releases
Nemotron 3.5 ASR is built for streaming multilingual speech recognition and voice agents. One 0.6B checkpoint. 40 language-locales. Sub-100m…
Nemotron 3.5 ASR is built for streaming multilingual speech recognition and voice agents. One 0.6B checkpoint. 40 language-locales. Sub-100ms latency. Cache-aware FastConformer carries context forward
Nemotron 3.5 ASR is built for streaming multilingual speech recognition and voice agents. One 0.6B checkpoint. 40 language-locales. Sub-100ms latency. Cache-aware FastConformer carries context forward between chunks instead of repeatedly reprocessing overlapping audio. Try it: http://www.together.ai/models/nvidia-nemotron-35-asr
Source: Together AI (X) | 2026-06-04