Benchmark

Accuracy you can check, not just claim.

Most legal AI shows you a score with no method behind it. We publish two, each with its method, its number, and the cases we still miss. The first measures whether our review catches real contract defects; the second measures retrieval, the foundation underneath it. Both are dated and reproducible.

96%
Review: planted defects caught
94%
Review: must-catch defects
80/100
Retrieval hit-rate
0
Fabricated citations, it abstains

Review accuracy

Run: July 2026 · 50 contracts

100%
Mexico
98%
Colombia
95%
Panama
90%
Brazil

Method

We built 50 contracts across Panama, Colombia, Mexico, and Brazil, each seeded with known defects and paired with a counsel-written answer key. The production review engine, the same model, prompts, and corpus grounding the product uses, reviews each one, and an independent model scores its findings against the key. We caught 96% of planted defects (291/304) and 94% of the must-catch ones (152/162).

Known weak point

Brazil is the weak point at 90%: its corpus is not built yet, so jurisdiction-specific defects still slip through, a missing BACEN registration of foreign capital, an absent LGPD data clause. Across the suite the review also raised 56 false positives against documented traps. Both are published here on purpose; the honesty is the point.

Retrieval benchmark (reproducible)

Run: July 2026 · 12 questions

CaseOn-codeCorrect article
Daños y perjuicios (CCF)83%
Pena convencional, límite (CCF)50%
Promesa de compraventa (CCF)100%
Constitución de sociedad anónima (LGSM)62%
Facultades de comprobación fiscal (CFF)100%
Régimen conyugal (CCF)100%
Testamento público abierto (CCF)100%
Zona restringida, sociedad (LIE)75%
Aviso a la SRE (LIE)25%
Fideicomiso zona restringida (LIE)38%
Renuncia al saneamiento (QRoo)100%
Obligaciones del vendedor (QRoo)100%

Method

For each question we know the code and the specific article the answer must come from. We measure two things: code correctness (are the retrieved sources from the right body of law, e.g. federal vs. a specific state) and article hit-rate (did the exact governing article appear in the top results). We report the honest number and the cases we miss, so the weak spots are visible and fixable.

What this number is, and isn't

This measures retrieval, the foundation. It is not inflated to look good; the 'pena convencional' case is a real miss we are tuning. Retrieval is only the floor: on top of it, every answer is synthesized under a strict-grounding prompt, its citations are structurally verified, an independent model checks that each claim is supported, and the system abstains when the corpus is not enough. That verification layer, not the retrieval score alone, is why an MXLEX answer is safe to rely on.

Read how verification works
MXLEX · Benchmark