Nemotron 3 Nano 30B A3B

Nemotron 3 Nano 30B A3B is an NVIDIA open-weight sparse MoE language model with 31.6B total parameters and approximately 3.2B active parameters per forward pass, designed for reasoning, chat, tool use, and agentic workflows. It features a hybrid Mamba-Transformer architecture with a 1M-token context window, targeting developers building chatbots, RAG systems, and AI agents. Source: supabase.