Legal Research Bench

As of Nov 3, 2026, Legal Research Bench (v1) is maintained by Vals AI. Metric accuracy (%). Ceiling 100. Saturation unknown. Contamination risk low.

Maintainer

Vals AI

Metric

accuracy (%)

Ceiling

100

Saturation

unknown

Contamination

low

Cadence

Declared cadence: as Vals updates (checked daily). Catalog snapshots: 1, latest Nov 3, 2026.

Source

Vals AI · this benchmark on Vals AI

Grading

AI judge (LLM-graded rubric) — All-pass on lawyer-written rubrics; every rubric item is checked by an LLM judge.

Partner-built

Fisher Phillips, McDermott Will & Schulte, Reed Smith, Legal Technology Hub

Source: Vals AI. A proprietary private test set run by Vals; scores are republished with permission. Rows are dated the day isitsota first saw them, and only the benchmark's standard harness counts.