Migration notice: Previously published versioned URLs are no longer maintained and do not redirect. Legacy V0 moved from /results/v1/ to /results/v0/, while current V1 moved from /results/v2/ to /results/v1/. Update saved links and integrations.
Legacy V0 — broader coverage
Legacy V0 with broader historical model coverage. Available benchmark scores are equally weighted; missing scores are excluded rather than treated as zero.
37 model-and-harness rows · data updated 2026-10-09
Current V1 — newer, narrower coverage
Current V1: clean-v1 reruns and updated evaluation sets, currently tested on fewer frontier models. The overall view requires complete coverage and uses configured evaluation-count weights.
38 complete-coverage model-and-harness rows · data updated 2026-10-09
Source/provenance: V2 score arrays in src/App.tsx, checked against the per-evaluation files in public/data/v1/. BioSecBench-Refusal scores come from public/data/v1/refusalbench.json. Overall weights come from public/data/aggregate-leaderboard.json. OncologyBench strict results use the selected 25 evaluations in the October 2 export, nine of them re-graded against a new decision rubric in the October 9 export, generated by scripts/build-oncologybench.mjs.