Observed arrival · 2026-09-07
ModelRigs: A Configurator for Local AI Hardware
ModelRigs compares machines for running open-weight AI models locally, factoring in memory fit, measured speed, electricity, tax, and ownership cost.
Field notes
The configurator combines model selection with machine selection, country, and buyer status, then distinguishes fit, throughput, and ownership-cost considerations. The homepage lists 21 hardware configurations, 27 open-weight models, 96 quantized variants, and 15 sourced benchmarks. Its methodology separates measured figures from estimates, assumptions, and official specifications, and says that disagreements between four sources are presented as a spread rather than collapsed into one average.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue