Your task, three models, side by side.
Benchmarks rank models on someone else's tasks. This is a bake-off on yours: pick a skill (or write a free prompt), give a brief, and watch three models answer at once. Pick the winner β the choice is remembered.