NVIDIA Nemotron 3.5 Content Safety: Customizable Multimodal Safety for Enterprise AI

Summary

NVIDIA released Nemotron 3.5 Content Safety on June 4, 2026 — a 4B-parameter multimodal safety classification model built on Google Gemma 3 4B that unifies text and image evaluation, multilingual coverage across 140+ languages, custom enterprise policy enforcement, and optional reasoning traces into a single inference call. The model evaluates user prompts, images, and assistant responses together in one context window, catching policy violations that emerge only from text-image interaction or r