Qwen-30B-A3B-NVFP4 NlM for DGX Spark Overview

Description:

Qwen3 is the latest generation of Qwen LLMs across dense and MoE variants, uniquely switching between “thinking” mode for complex reasoning and “non‑thinking” mode for fast dialogue. It advances state‑of‑the‑art performance in math, coding, and commonsense reasoning, improves alignment for creative writing, role‑play, and multi‑turn instructions, integrates precisely with external tools, and supports 100+ languages. The Qwen3‑30B‑A3B MoE model is a pretrained and post‑trained causal LM with 30.5B total parameters (3.3B active; 128 experts with 8 active) and a 32k context window (up to 131k with YaRN).