Skip to content
K = 0.731
Benchmark sources

Rapidata SVG · Human Preference

Rapidata · People compare model-generated vector graphics to evaluate visual preferences directly.
Official benchmark
Across e/accUnder observation
Evidence budgetNot scored
Upstream data timeUnverified
Last successful syncCollection not started

What it measures and how

Use only Preference Elo for the same static SVG task. Alignment, Coherence and Overall add no duplicate votes. Use public score tables from a fixed official dataset version, traceable to original human comparisons.

How this evidence is used

Under observation: official data and licensing are available, but three rolling DeepSeek aliases lack fixed-version evidence, so the entire score table is withheld. Awaiting run versions from the organization; active sources’ budgets have not been reduced.

Limitations and attribution

SVG drawing does not represent all interface design. Public raw data lacks model versions, generation dates and API fingerprints, so rolling API version changes cannot be established. A leaderboard update date cannot establish a model version.

Data license: CC BY 4.0 · Rapidata/svg-benchmark; retain official attribution and data links.

Results are published by Rapidata. Raw scores and the e/acc consensus index use different scales and cannot be added directly.