31.6B / 3.6B (2 variants: Base
NVIDIA Open Model License
Instruct; hybrid Mamba-Transformer MoE),Breakthrough agentic AI model with hybrid Mamba-2 + Transformer + MoE architecture (activates 6 of 128 experts per pass); 1M-token context window natively; both Base and Instruct (post-trained) variants released; 4x faster throughput than Nemotron 2 Nano; 3.3x higher throughput than Qwen3-30B-A3B and 2.2x vs GPT-OSS-20B on H200; reduces reasoning token generation by 60%; excels at math, coding, multi-step tool calling, and multi-turn agentic workflows; reasoning ON/OFF modes with thinking budget controls; trained with multi-environment RL using NeMo Gym; includes 3T new training tokens openly released; designed for multi-agent systems at scale; runs on RTX GPUs to data center; Super (100B/10B) and Ultra (500B/50B) variants coming H1 2026.