PRBench: Large-Scale Expert Rubrics for Evaluating High-Stakes Professional Reasoning

TL;DR

PRBench is a large, open-source benchmark with expert-curated tasks in Finance and Law, designed to evaluate AI models on real-world, high-stakes professional reasoning, revealing significant gaps in current model capabilities.

Contribution

The paper introduces PRBench, the largest public rubric-based benchmark for legal and finance domains, with expert-authored tasks and rigorous validation, enabling detailed assessment of AI performance in professional contexts.

Findings

Top model scores are only around 0.39 and 0.37, indicating substantial room for improvement.

Models often make inaccurate judgments and lack transparent reasoning.

Performance varies significantly across different capabilities and failure modes.

Abstract

Frontier model progress is often measured by academic benchmarks, which offer a limited view of performance in real-world professional contexts. Existing evaluations often fail to assess open-ended, economically consequential tasks in high-stakes domains like Legal and Finance, where practical returns are paramount. To address this, we introduce Professional Reasoning Bench (PRBench), a realistic, open-ended, and difficult benchmark of real-world problems in Finance and Law. We open-source its 1,100 expert-authored tasks and 19,356 expert-curated criteria, making it, to our knowledge, the largest public, rubric-based benchmark for both legal and finance domains. We recruit 182 qualified professionals, holding JDs, CFAs, or 6+ years of experience, who contributed tasks inspired by their actual workflows. This process yields significant diversity, with tasks spanning 114 countries and 47 US jurisdictions. Our expert-curated rubrics are validated through a rigorous quality pipeline, including independent expert validation. Subsequent evaluation of 20 leading models reveals substantial room for improvement, with top scores of only 0.39 (Finance) and 0.37 (Legal) on our Hard subsets. We further catalog associated economic impacts of the prompts and analyze performance using human-annotated rubric categories. Our analysis shows that models with similar overall scores can diverge significantly on specific capabilities. Common failure modes include inaccurate judgments, a lack of process transparency and incomplete reasoning, highlighting critical gaps in their reliability for professional adoption.