Nvidia launched Nemotron 3.5, a multimodal safety model designed to detect harmful content across text and images. It allows enterprises to customize safety guardrails for specific regional or industry requirements. This flexibility reduces the friction of deploying LLMs in regulated markets. Developers can now tune safety thresholds without retraining the core model.