Estimates, not facts
Every score, rank, and mention rate on this site is a snapshot of how ChatGPT responded under the conditions we tested. It is not a definitive statement of who is “best,” who will win a deal, or how any single user will be answered.
Results depend on the situation
Visibility shifts with prompt wording, category framing, timing, model version, web-search behavior, geography, and what else the model has seen recently. Change any of those and the list can change. Treat what you see here as directional: useful for comparison, not as a fixed truth.
AI answers are non-deterministic
Large language models do not return identical answers every time. The same question can yield different brands, orderings, or omissions across runs. Our weekly samples reduce noise; they do not eliminate it. Week-to-week movement can reflect model variability as much as real market change.
Verify before you act
Check and verify anything material before you rely on it, whether for reporting, competitive claims, investment decisions, or customer promises. Run your own prompts. Cross-check with other sources. This index is meant for estimates and research only. Your situation may vary, and outcomes can differ.
