NVIDIA Nemotron 3 Super: 120B Open Model With Only 12B Active — Built To Be An Agent's Brain

NVIDIA's Nemotron 3 Super is a 120B-parameter, 12B-active hybrid Mamba-Transformer MoE that scores 60.47 on SWE-Bench Verified, holds 91.75 on RULER at 1M tokens (while GPT-OSS-120B collapses to 22.3), and ships fully open — weights, datasets, and training recipes.