Event ·Multi-source

同一任务对比:Claude Opus 5.5 用 3270 次动作,GPT-6 Astra 只用 514 次

Artificial Analysis 的 AA-Robotics v0.1 测试让两个模型在机器人上执行同一任务:手写 hello world。Claude Opus 5.5 采取逐点描画方式,共用了 3270 次手臂移动,耗时 34 分钟。GPT-6 Astra 采用连续笔画,只需 514 次移动,13 分钟完成。视频从模型视角展示了两种截然不同的书写策略。

Current read

两个大模型在机器人上写同一个 hello world,一个磨了 34 分钟用 3270 次动作,另一个 13 分钟搞定,差异很直观。

11 sources64° AI heatLast updated 2026/10/28 22:12:18

Evidence

11 sources
Primary sourceArtificial Analysis
View source
Related source 1kimmonismus
View source
Related source 2TestingCatalog
View source
Related source 3@koltregaskes
View source
Related source 5Yuchen Jin
View source
Related source 7rohanpaul_ai
View source
Related source 10Simon Willison’s Weblog
View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.