Anthropic 内部新模型被 METR 称为 Model 4，全部基准首破 50%
精选理由
METR 透露 Anthropic 上月训完的新模型，全部测试基准首次都过 50%，是这批模型里分最高的。
Anthropic 一个上个月完成训练的内部新模型，被评测机构 METR 标记为 Model 4（Anthropic 另有内部代号）。在 METR 分析的模型集合中，它是得分最高的一款。它也是首个在其受测的每一项基准上都超过 50% 的模型。
原文 · Andrew Curran
Anthropic has a new internal model that finished training last month. METR refers to it as Model 4 (Anthropic has its own internal codename). It is the highest-scoring model in the set METR analyzed, and it is the first to clear 50% on every benchmark it was tested on. https://t.co/2mcv9Oijka
