StagingBench is an independent benchmark of AI virtual staging tools. It stages the same 10 empty room photographs in each tool using the modern style and default settings, then presents the outputs side by side. The rooms include open plan layouts, sloped attic windows, bay windows, small spaces, and a room that is not fully empty. The source photographs are CC0 or public domain. The benchmark uses a fixed, documented script to check results for changes to walls, windows, and doors, as well as crop, zoom, and framing shifts. It also records details such as output resolution, time to first result, furniture scale, lighting, and shadows. At the time described, one editor provides realism ratings by comparing each output with the original; these ratings are not blind, and blind ratings from agents and photographers are planned. Testing is partial, so provisional comparisons are based on the rooms and tools tested so far rather than a final overall score. Users can compare staging results, review automatic checks, and consult information drawn from vendors’ public pricing and feature pages. The workflow gives each tool identical inputs and one attempt, without prompt adjustments or re-rolls, and distinguishes verified information from details that could not be confirmed. StagingBench is a comparison and testing resource, not a virtual staging tool for editing users’ own photographs.