RTX 3050

1500 Yuan GPU Runs Qwen 30B — Local LLM Hardware Bar Drops Again

Reddit user runs Qwen 30B MoE on RTX 3050 6GB, hitting 30 tokens/sec at 90k context. Enterprise IT should reassess local LLM costs.