Put two to four models side by side across price, context, capabilities, benchmarks, performance and observed usage.
Select 2–4 models. The URL updates as you change the selection.