Qwen 3.6 35 UD 2 K_XL is pulling beyond its weight and quantization (No one is GPU Poor now)
Hi guys, Back again. I have tested the Qwen 3.6 UD 2 K_XL Unsloth model on the same paper to web app task. The model is performing very well. It handled all tool calls properly and also managed large context using llama.cpp on a 16GB VRAM on lapt…