A model column is one AI model's independent ranking of a category, displayed side by side with the other models' — the format that makes cross-model disagreement visible.
Averaging models into a single list hides the most useful signal: where they disagree. A model-column layout keeps each model's ordering intact — ChatGPT's list next to Claude's next to Gemini's next to Perplexity's — so a brand's spread is visible at a glance.
Reading columns is how practitioners localize a problem: strong on three models and weak on one points to that model's sources; weak everywhere points to the category story itself.