Nemotron 3.5 Lightning
Nemotron 3.5 Lightning is an open 30B Mixture-of-Experts model with 3B active parameters, supporting a 1M-token context and commercial use, pretrained on 20T+ tokens (data cutoff Sep 2025). Built for high-volume, low-latency agentic workloads like planning and tool selection, it runs via Ollama, Hugging Face, and NVIDIA's platform.