To solve this “utility problem,” researchers have introduced a rigorous new testing ground called Agents’ Last Exam (ALE). The name carries a dual meaning: it acts as a final graduation exam to prove an AI agent is actually ready for corporate deployment, and it represents the absolute frontier of what today’s technology can handle.

The creators of ALE don’t intend for it to be a static, one-time leaderboard. Designed as a “living benchmark,” its pool of tests will continuously grow as new industries and workflows evolve. Ultimately, the goal of Agents’ Last Exam is to shift the AI industry’s focus away from winning abstract academic trophies and toward creating digital assistants capable of driving genuine, measurable economic growth.

Challenge and measure AI agents on economically valuable and real-world tasks.

Agents’ Last Exam is building the largest-scale, broadest-coverage agent evaluation benchmark to date, measuring performance on long-horizon, economically valuable tasks with verifiable outcomes. Led by Berkeley RDI and 300+ industry experts, it now spans all 55 targeted sub-industries covering most major fields of professional work performed on a computer, with 1,500+ tasks collected toward a 5,000-task target, keeping scores objective, comparable, and meaningful across domains.