125 tok/s for Qwen3.6 q4xl on 2x 4060ti is insane perf/dollar
Signal
45
Hype
35
In three linesUser reports 125 tokens/s with Qwen 3.6 Q4 quantized on 2x RTX 4060 Ti (~$1000, 32GB VRAM). Outperforms high-end 2026 mini-PCs at fraction of cost. Testing CUDA 13.3 optimization to reach 150 tok/s.Read source
Your take?
Summary generated by Claude — human-verified