Independent Benchmark Validates Alexi’s Performance in the 2025 Vals Legal AI Report

The 2025 Vals Legal AI Report (VLAIR), one of the most comprehensive independent evaluations of legal AI systems to date, has confirmed Alexi’s strong performance across accuracy, citation authority, speed, and reliability, establishing it as one of the most dependable enterprise-ready AI platforms for legal research and workflow automation.

Notably, many well-known legal technology companies chose not to participate in this year’s evaluation. That makes Alexi’s decision to undergo, and excel in, an independent, data-driven audit especially significant at a time when transparency and accountability in legal AI are under increased scrutiny.

Why This Matters

Since May, reported cases of AI-generated “hallucinations” in legal filings have surged by more than 250%, putting renewed pressure on legal technology providers to prove their reliability. The VLAIR study was designed precisely for this reason: to move beyond marketing claims and establish objective performance standards for AI systems used in law.

Developed in collaboration with Consortium Firms and legal experts, the report tested leading AI platforms on realistic research tasks such as interpreting statutes, summarizing case law, and analyzing multi-jurisdictional issues. Its dataset included 200 question-and-answer pairs referencing U.S. federal and state laws, regulations, and judgments, making it one of the largest and most rigorous assessments of its kind.

Key Findings

The VLAIR evaluation found that Alexi performed at or near the top of every measured category:

• 79% single-jurisdiction accuracy, tied for the highest score among all systems tested.
• 80% overall accuracy, compared to a 71% lawyer baseline (a 13% relative improvement).
• 8% higher rate of valid primary law citations than ChatGPT, reducing compliance and malpractice risk.
• Second-fastest response speed while maintaining a 99% response rate, in contrast to ChatGPT’s four-times-higher non-response rate.

Together, these results highlight a platform that combines legal accuracy, authoritative sourcing, and operational reliability, a balance many general-purpose AI models have struggled to achieve.

What It Means for Legal Teams

In practice, these results translate to fewer revisions, lower compliance risk, and more efficient workflows. Lawyers spend less time correcting AI outputs and more time applying legal judgment.

As Mark Doble, CEO of Alexi, put it:

“These results confirm what our customers experience every day: Alexi delivers the superior legal accuracy and compliance leading law firms need from AI, without compromising on speed or reliability.”

For firms weighing which tools to trust with substantive legal research and drafting, independent validation like this provides tangible assurance that performance claims are backed by evidence.

The Broader Context

While the generative AI field evolves quickly, the legal profession’s threshold for reliability remains high. Many vendors continue to rely on internal testing or anecdotal success stories. Alexi’s participation in an external, data-driven benchmark, particularly when some competitors opted out, signals a commitment to transparency, accountability, and measurable trustworthiness.

The findings also underscore a broader shift: as legal AI becomes more deeply integrated into daily workflows, accuracy, citation integrity, and consistency are becoming the defining metrics of adoption, not novelty or hype.

Methodology in Brief

The Vals Legal AI Report evaluated AI systems on their ability to answer complex questions grounded in U.S. law. Each response was assessed for legal reasoning, citation validity, and jurisdictional precision. Tasks included confirming definitions, interpreting fact patterns, identifying relevant statutes and precedents, and conducting comparative analyses across jurisdictions.

The resulting benchmark represents one of the most transparent and comprehensive efforts yet to measure how generative AI performs in real legal practice.

In a year when many chose not to be tested, Alexi did and excelled.
Independent validation now confirms what many firms already experience in practice: Alexi’s legal accuracy, reliability, and compliance readiness make it one of the most capable AI systems operating in law today.

‍

View our VALS application report

View the VALS Legal Research report