Model by nvidia

nemotron-3-super-120b-a12b

Open, efficient hybrid Mamba-Transformer MoE with 1M context, excelling in agentic reasoning, coding, planning, tool calling, and more