MODEL BAKEOFF · HUMAN REVIEW

Which dashboard would you want to use?

Inspect every entry using the same scenarios. Rank from 1 (best) to N (least preferred). Models, automated scores, tokens and timings stay hidden until you submit.

Your ranking

Consider visual clarity, information hierarchy, usability, responsive layout and whether this feels useful in your homelab. Use the common controls above and interact with search, filters and service details. Unavailable entries remain in the ranking.

Rankings are kept only in this browser. Export downloads a JSON file; attach it to the DSH conversation to freeze your ranking and reveal the model mapping. Nothing is submitted automatically.