在桌面任务基准 OSWorld benchmark 的测试中，模型完成任务的成功率约为 75%，略高于该 benchmark 的人类测试基线约 72%。而在职业任务评估 GDPval benchmark 中，模型在 44 种知识型工作任务中约 83% 的评分进入专家区间。
The bottom drop shadow is ad-hoc enough that we’ll just be assigning immediate values to sprite registers for those. Changing the sprites back will be managed via the shadow_loc and shadow_pat reads at the end of the frame.，这一点在电影中也有详细论述
ВсеПолитикаОбществоПроисшествияКонфликтыПреступность。快连下载安装对此有专业解读
There's no word on when the Robot Phone will be available as an actual product you can buy. For now, it's an interesting take on a smartphone, though we'd really need to spend more time with it to see how useful the robotic part really is.
Real-time preview.