

I run a “server” (old 9-year gamer laptop) with an 8 GB Geforce 1070 Max-Q. It’s currently running full GPU-offloaded stock llama.cpp, with Unsloth’s Qwen3.5 9B at UD Q4_K_XL (good enough quantization for coding assistance and short agentic tasks) and a 90k context at around 300 t/s prefill and 25 t/s decode, which are good enough speeds for daily use.
It can’t vibecode GTA6, but it can help with quick scripting, research and things like that.





Opens up? It’s pretty much pure grind, level after level, unless you got very lucky with your class and skills, or got the “rich” trait at character creation.
Better go play a less bad game, like Diablo 4.