Repository Name

The selected name is local-llm-bench-lab.

Other names considered:

• apple-silicon-llm-bench: clear, but too tied to one hardware family.
• gemma4-bench-lab: accurate for the first experiment, but too narrow for future models.
• local-ai-bench-lab: broader, but weaker as an LLM-focused benchmark repository.

The goal is to compare local LLM runtimes and model setups in a repeatable way. local-llm-bench-lab keeps the scope model-independent while leaving room for setup notes, speed tuning, and future experiments.