Nemotron 3 Nano Omni is NVIDIA’s new open omni-modal model — unified video, audio, image, and text reasoning in a single efficient model built for agentic workloads.
-
Best-in-class on MMlongbench-Doc, WorldSense, DailyOmni, and VoiceBench leaderboards
-
~3× higher throughput and ~2.75× lower compute for video reasoning
-
30B-A3B Hybrid MoE — efficient enough for sub-agent roles at scale
-
Fully open: weights, datasets, and training recipes all released
-
Available on Hugging Face, NVIDIA NIM, and 17+ partner platforms
📚 Resources:
NVIDIA Technical Report :https://nvda.ws/48u8DvP
Hugging Face Model Page:https://nvda.ws/4ue1x7d
Tech Blog: https://nvda.ws/4vQXUFJ
Cookbooks: https://nvda.ws/4sYvVBi
Stay up to date on NVIDIA Nemotron by subscribing to NVIDIA news and following NVIDIA AI on LinkedIn, X, YouTube, and the Nemotron channel on Discord. Access open Nemotron models on Hugging Face and NIM microservices on build.nvidia.com. Share your ideas and vote on what matters to help shape the future of Nemotron.