Benchmarks

Our benchmarks review the metrics that matter when sizing a homelab: throughput (tokens/s for local LLMs, IOPS for storage), latency, power draw, RAM usage, temperatures. Most of our benchmarks are not based on measurements we took ourselves: cross-check their figures against manufacturer spec sheets and official documentation before relying on them. We cover GPUs, mini-PCs, NAS units, VPS and local AI models. No content is sponsored. Use these pages as a starting point to size your homelab or inference server, then head to our shop for the gear mentioned.