RTX 5060ti

Two 5060ti GPUs Force-Run a 100B-Param Qwen Locally — 'Barely Usable'

A Reddit user squeezed a quantized Qwen onto two RTX 5060ti GPUs at 30 chars/sec — local AI's practical tipping point. Runnable, but not useful.