I benchmarked quants of Qwen 3 .6b from q2-q8, here's the results:

r/LocalLLaMA
Open Source AI

Submitted by /u/PraxisOG [link] [comments]