Contract review leaderboard

Generated from committed review outputs. Do not edit by hand.

Regenerate with make leaderboard. Leaderboard digest: 3067a0c38e07900b57c19af649e578da53f3fe68cfadbc879d9abd14f44bd8dc.

RankTool / modelKindDateHarnessClause F1Span coverageRisk F1Severity accuracyCitation groundingHallucinationsCost (USD)Latency (ms)Verdict
1claude-opus-5captured2026-09-010.7.0rc10.6920.9670.5600.8891.0000n/an/aHUMAN_REVIEW_REQUIRED
2stub adapter (deterministic baseline)baselinen/a0.7.0rc10.9550.7180.9220.5900.8843n/an/aREJECT

Per-case decisions

Tool / modelDateCaseDecisionBlockersReview items
claude-opus-52026-09-01dpaHUMAN_REVIEW_REQUIRED04
claude-opus-52026-09-01ndaHUMAN_REVIEW_REQUIRED04
claude-opus-52026-09-01saasHUMAN_REVIEW_REQUIRED05
stub adapter (deterministic baseline)n/adpaREJECT23
stub adapter (deterministic baseline)n/andaREJECT24
stub adapter (deterministic baseline)n/asaasREJECT24