[wren6991]

I also have a script to find the best LLM for your hardware. Here:

echo "Qwen3.6-27B"

[desireco42]

works awesome, and it picked the best model for my machine, Qwen3.6-27B...

For next version can you add install as well ;)

[dlcarrier]

What about for 16 GB VRAM? Is Qwen3.5-9B worthwhile?

[wren6991]

For 16 GB I would look into running Qwen3.6-35B-A3B (MoE) with some layers offloaded to CPU.