ITBench-AA:前沿模型在首个面向企业 IT 智能体任务的基准测试中得分低于 50% —— 由 Artificial Analysis 与 IBM AI资讯 · 9 阅读 · 2026-05-28 · 自动采集 ITBench-AA: Frontier Models Score Below 50% on the First Benchmark for Agentic Enterprise IT Tasks — by Artificial Analysis and IBM