英伟达发布Nemotron 3瞄准AI智能体

Group 1 - The core announcement is that Nvidia has launched a new open-source large language model called Nemotron 3 Super, specifically designed for enterprise-level multi-agent systems [2] - The new model features a novel mixture of experts (MoE) architecture, which enhances inference throughput to more than five times that of the previous generation model [2]