ITBench-AA:前沿模型在首个企业IT代理任务基准中得分低于50% —— Artificial Analysis与IBM AI资讯 · 12 阅读 · 2026-05-30 · 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