AI 考试全优，但一干活就废？新基准揭底
AI 在各类考试中拿高分，但放到真实工作里却频频掉链子。这篇论文认为，问题出在考试本身：现有基准测试考的是“做题”，不是“干活”。他们联合 250 多位行业专家，搞了一个叫 Agents' Last Exam 的新基准，覆盖 13 个行业、1000 多个真实工作任务——从写报告到做财务分析，全是能直接产生经济价值的那种。结果呢？最强 AI 的平均通过率只有 2.6%。换句话说，AI 离真正帮你干活还差得远。这个基准会持续更新，目的不是搞排行榜，而是逼着 AI 去解决那些能赚钱的问题。
📄 原文摘要(英文)
Recent AI systems have achieved strong results on a wide range of benchmarks, yet these gains have not translated into economically meaningful deployment across many professional domains. We argue that this gap is largely an evaluation problem: widely used benchmarks lack sustained performance measurement on real and economically valuable workflows. This paper introduces Agents' Last Exam (ALE), a benchmark designed to evaluate AI agents on long-horizon, economically valuable, real-world tasks with verifiable outcomes. Developed in collaboration with 250+ industry experts, ALE covers non-physical industries defined with reference to O*NET / SOC 2018 (the U.S. federal occupational taxonomy). It is organized around a task taxonomy with 55 subfields grouped into 13 industry clusters covering 1K+ tasks. Current results show that the hardest tier remains far from saturated: across mainstream harness and backbone configurations, the average full pass rate is 2.6%. ALE is designed as a living benchmark: its task pool grows continuously as new workflows and industries are onboarded. More broadly, ALE is intended not merely as another leaderboard, but as an instrument for closing the gap between benchmark success and GDP-relevant impact.