index / etsabary/visual-decision-benchmark
visual-decision-benchmark
200 image-grounded decisions: 21 complete routes (18 hosted decision models, two local RTX 2060 models and a Sol reasoning-off LLM reference), plus three LLM speed stops. Accuracy, latency, cost and 20 named categories.
Research and EvalsUnclearListedPython
Repository
Stars0
LanguagePython
LicenseMIT
Created2026-10-15
Last pushed2026-10-28
Officialno
Judgment
Genuine Jev project0.75
gate 0.5
Uses Jev at runtime0.59
Is a meta list0.07
Is a reimplementation0.09
Quality scales
Substance2.20 / 3
gate 0.5
Docs2.01 / 3
Novelty2.76 / 3
Composite score0.77
Category distribution
Research and Evals1.00
SDKs and Clients0.00
Integrations0.00
Agent and Dev Tooling0.00
Applications0.00
Games and Simulation0.00
Learning0.00
Other0.00
Assigned categoryResearch and Evals
Category probability1.00
Category confidence1.00
PatternUnclear
Pattern confidence0.19
Curation
StatusListed
Reasongate passed
README pickno
Overridenone
Sources
Provenance
Modeljev-1.14.0
Question setv2
Judged at2026-10-28 12:09 UTC