Discover NVIDIA Nemotron 3.5 Lightning, an open 30B Mixture-of-Experts model with 3B active parameters built for fast, low-latency execution in long-running AI agents. Learn about its speculative decoding, NVFP4 quantization, NeMo Switchyard routing, and key performance benchmarks.
