replaced MiniCPM5-1B with MiniCPM4.1-8B

Fix q8 quant

disabled reasoning; changed default quant from q4 to q8

updated model from MiniCPM3 to MiniCPM5

Fix ZeroGPU: preload CUDA runtime libs for the llama.cpp wheel

Fix Python 3.10 startup: tomllib fallback + pin python_version 3.12

Run a local small model (MiniCPM3-4B via llama.cpp) instead of a cloud API

Initial Gradio Space