Nvidia introduced Nemotron 3.5 Lightning, a 30B-parameter open mixture-of-experts model, and Switchyard, a router that dynamically...

Source Brief

Nvidia introduced Nemotron 3.5 Lightning, a 30B-parameter open mixture-of-experts model, and Switchyard, a router that dynamically assigns tasks to appropriate models mid-task, cutting task costs by up to a third in Nvidia's tests. This addresses the cost tradeoff in running always-on AI agents.