Nemotron 3 Nano Omni

Released on April 28, 2026

Changes from previous version

Adds native image, video, and audio understanding to the family after Nemotron 3 Super, trading its 1M text context for a 256K omni-modal context optimized for perception sub-agents.

Release Summary

NVIDIA's open 30B-total, 3B-active omni model combines reasoning over text, images, video, and audio for document intelligence, computer use, and multimodal agents. NVIDIA released the model with open weights, training data, and recipes, plus a 256K context window and 64K maximum output.

Timeline

April 28, 2026

Nemotron 3 Nano Omni released

NVIDIA releases an open omni-modal reasoning model for unified text, vision, audio, and video agent workflows.