Qwen3.6 27B on dual RTX 5060 Ti 16GB with vLLM: ~60 tok/s, 204k context working
I’ve been testing Qwen3.6 27B on a pretty non-standard local setup and figured the numbers might be useful for anyone looking at the newer 16GB Blackwell cards. Hardware: 2x RTX 5060 Ti 16GB 32GB total VRAM Proxmox LXC 16 vCPU ~60GB RAM CUDA 13 / Torc…