NVIDIA announced an expansion of its Nemotron 3 model family with the introduction of Nemotron 3.5 Lightning, designed for long-running AI agent workloads and touted as the most efficient of its kind. Nemotron 3.5 Lightning is a 30-billion-parameter Mixture-of-Experts model that can increase output speed by up to 4x, accelerating AI agent task completion by 30%. Concurrently, NVIDIA also released NeMo Switchyard, an open-source library for intelligent routing in AI agent tools, which, according to internal benchmarks, can reduce task completion costs by nearly two-thirds while maintaining state-of-the-art accuracy.