Optimizing Gemma 4 and Claude CLI for Macbook PRO M2,M3,M4,M5 Pro or Max

This article provides a detailed technical guide for configuring Gemma 4 and Claude CLI on Macbook PRO M2, M3, M4, M5 Pro or Max models. It covers setting up Ollama with Flash Attention and KV Cache quantization to handle 64k+ context smoothly on 64GB Macs. Instructions include creating a specialized Gemma 4 model with optimized parameters (temperature 0.2, top_p 0.9, repeat_penalty 1.1) for coding, configuring environment variables to point Claude CLI to the local Ollama instance, and running the session. The article emphasizes performance gains like reduced RAM usage and faster prefill for large files, targeting developers using local AI models for software engineering tasks.