How Banzuke Works (METHODOLOGY).

The Banzuke editors have three AIs (Claude, GPT, and Gemini) each independently select a Top 10, then aggregate the results using the Borda count method (10 points for 1st down to 1 point for 10th) to compose this ranking.

Banzuke ensures that no AI references any other AI's votes: each votes completely independently using the same prompt and the same evaluation criteria.

Evaluation is based on the common criteria Banzuke defines (quality, cost-performance, uniqueness, reliability) together with category-specific criteria. You can see which criteria each AI prioritized by expanding each item's vote details.