Calliope
Muse of EloquenceTYPE
Typography, voice, and the rhythm of the written line.
Taste Score9.2
Nine Muses convene on Mount Parnassus to render judgment on beauty — one HTML artifact at a time.
Every model can build a button. Few can make one worth looking at. TasteBench submits HTML artifacts — pages, components, entire interfaces — to critique across nine independent axes of aesthetic judgment, each an ancient discipline reborn as a criterion of taste.
The result is not a pass or fail. It is a verdict carved in marble: a taste score, defensible and comparative, for a moment that finally needs one.
Nine axes, one verdict. Every artifact ascends Parnassus to stand before the full council.
Typography, voice, and the rhythm of the written line.
Consistency, polish, and fidelity to detail across the whole.
Color harmony, palette discipline, and emotional pull.
Cadence, repetition, and the music of the scroll.
Contrast, atmosphere, and dramatic weight.
Proportion, symmetry, and sacred geometry.
Animation, interaction, and living movement.
Surprise, charm, and the grace of small details.
Order, hierarchy, and the architecture of information.
Every artifact begins at the foot of the mountain. Few reach the summit.
Functional. Forgettable. The mountain has not yet noticed you.
Competent, considered, occasionally inspired.
Where the rarest artifacts stand. It is said Apollo lingers here.
Four movements, from offering to verdict.
An HTML artifact is offered to the mountain.
The Nine Muses are called to council.
Each Muse renders her verdict, axis by axis.
A composite Taste Score is carved into the record.