NVIDIA Nemotron 3 Nano Omni

NVIDIA

An open multimodal reasoning model designed to unify vision, audio, and language processing for highly efficient AI agents. It enables faster, smarter responses across diverse input types with significantly higher throughput than previous open models.