Anthropic Fable 5.2 真实自由职业任务完成率达 41%
原标题:Anthropic's Fable 5.2 completes 41 percent of real freelance jobs, up from 2.5 percent a year ago
AI 摘要
Anthropic 的新模型 Fable 5.2 在 Scale Labs 与 CAIS 的 Remote Labor Index 基准上完成 41.3% 的真实自由职业项目,远超一年前 Manus 的 2.5% 和今年 7 月 Fable 5 的约 16%。该基准包含 240 个来自 Upwork 的真实项目,总价值约 14.4 万美元,由人类专家在专业软件中人工验收。Fable 5.2 在音频编辑、建筑与室内设计、3D 建模上进步最大,但仍难以处理工业产品设计和动画。
正文节选
Anthropic's Fable 5.2 completes 41 percent of real freelance jobs, up from 2.5 percent a year ago Anthropic's new model Fable 5.2 makes a massive leap on the Remote Labor Index, completing more than 41 percent of real freelance projects at professional quality. The system leaves both its direct predecessor and the entire competition far behind. On the Remote Labor Index (RLI) from Scale Labs and the Center for AI Safety (CAIS), Anthropic's new model Fable 5.2 takes the lead with 41.3 percent. Th