Nasir

Hardware: MacBook Pro 4 (M4 processor) with 48 GB RAM, typically 36 GB allocated to VM for LM Studio/MLX/local inference and optimal model performance. Preferences: High-performance Apple Silicon-optimized models, particularly quantized formats (GGUF/MLX). Focus on models that run efficiently on Apple Silicon with MLX framework optimization. Prefer 4-bit and 8-bit quantized models for optimal memory usage within the 36 GB VM allocation. Local Apps: Jan, MLX LM, LM Studio, Ollama - all configured for MacBook Pro 4 hardware specifications.