Guides
Plain-English explainers on the parts of running local LLMs that matter most.
- Local model performance: what makes an LLM fast or slow
Why memory bandwidth decides tokens per second, the levers you control, and the speeds people report on each GPU and Mac.
- How to compare local LLM benchmarks
Check GPU count, quantization, context and concurrency before comparing tokens per second.
- Quantization for local LLMs
What Q4_K_M, Q5, MLX-4bit, and the rest actually mean — and how to pick the right one for your hardware.
- VRAM requirements for local LLMs
How much VRAM you actually need, by model size and quantization level. Includes a what-fits-by-tier reference table.