Comment by verdverm

You won't need a frontier size model for most tasks before long. Qwen 3.6 (small) punches way above its weight. I run it at home @8bit on an OEM Spark