SemiAnalysis 发布 GDPx 真实工作基准，AI 智能体最高仅得 12%
原标题：No AI agent scores above 12 percent on SemiAnalysis' new real-world work benchmark
AI 摘要
分析机构 SemiAnalysis 发布 GDPx 基准，用 206 个跨 11 个领域的真实工作任务测试 AI 智能体，任务需在专业软件中持续数小时。结果显示所有模型得分均不超过 12%，Claude Opus 5.5 以 12% 领先，Gemini 4 Carbon 以 10.9% 紧随且成本约为三分之一，GPT-6 Astra 为 9.6%。批评者指出安全过滤器扭曲排名、测试任务不公开、各模型使用自家智能体软件导致条件不平等，开发者 Josh Purtell 称其“噪音多于信号”。
正文节选
No AI agent scores above 12 percent on SemiAnalysis' new real-world work benchmark Key Points - Analytics firm SemiAnalysis has released GDPx, a benchmark that tests AI agents on 206 tasks in realistic work environments and professional software over several hours. - The scores are low. Claude Opus 5.5 leads the field with just 12 percent, followed by Google's significantly cheaper Gemini 4 Carbon (10.9 percent) and GPT-6 Astra (9.6 percent). - Critics flag distortions from safety filters, the s