AI Model Reviewer

Categories

Every output grouped by what it depicts. All pelicans sit under Pelican, all controllers under Game controller, and so on. Each page shows our runs first, then outputs from an older, separate SVG benchmark.

Legacy · different prompt Legacy outputs answered different prompts, so they are shown for reference only: they are never scored, ranked, priced or compared with our runs, and no date or cost exists for them. The whole legacy SVG bench (1,113 outputs shown; 218 withheld: provenance unclear).

38 categories · 596 of our runs · 1,113 legacy SVGs

Our runs and legacy outputs

Subjects covered by both our tests and the legacy SVG bench.

Our tests only

Legacy SVG bench only

Legacy prompts with no outputs yet

6 categories