A taste for tradeoffs.
How much soup does another dollar buy?
Quality & efficiency
Loading evaluations…
Models
The tasting hasn’t started yet.
Existing bowls are waiting for their first visual evaluation.
Up and to the right is better: higher quality, lower cost or time.
The tasting notes
Scores range from 0 to 1. Click a row to expand its judge notes.
Standalone HTML is a hard gate: external dependencies make the composite zero. Otherwise, the six visual criteria and the three judges have equal weight. Each judge’s samples are averaged first. Vegan scores describe visible ingredients; they cannot establish a recipe. Browser frame timings are diagnostics and do not affect reward.
| Model / effort | Reward | Cost | Minutes | Standalone | Vegan | Appetizing | Steam | Aesthetics | Interaction | Reduced motion |
|---|