Nvidia Launches Nemotron 3 Super with Advanced AI Capabilit

Nvidia has unveiled Nemotron 3 Super, a groundbreaking 120-billion-parameter hybrid AI model designed to address the inefficiencies of traditional architectures in enterprise tasks. This release combines state-space models, transformers, and a novel Latent Mixture-of-Experts (LatentMoE) design, allowing for improved context handling and processing speed, all while remaining available under mostly open weights for commercial use. The model is optimized for the Nvidia Blackwell GPU platform, delivering up to 4x faster inference than previous architectures without sacrificing accuracy.

The introduction of Nemotron 3 Super marks a significant technical shift in the AI landscape, providing enterprises with a powerful tool for complex workflows involving multi-step reasoning and large document analysis. Its enhancements in memory efficiency and token processing demonstrate Nvidia's commitment to pushing the boundaries of AI capabilities, effectively increasing national AI autonomy. By enabling quicker and more accurate model performance for critical applications, Nemotron 3 Super positions Nvidia at the forefront of sovereign AI solutions, fostering reduced dependency on foreign tech for enterprise AI tasks.