GPT-6 Astra vs MolmoAct2 on bimanual robotic manipulation. 5 tasks, 200 graded trials. Astra 7/100 completions vs MolmoAct2 0/100; mean progress 46 vs 12 out of 100.

In a new robotics benchmark, GPT-6 Astra shows major gains in spatial understanding. On StationeryBench, the model completed 7 out of 100 tasks with dual-arm robots, while…

GPT-6 Astra vs MolmoAct2 on bimanual robotic manipulation. 5 tasks, 200 graded trials. Astra 7/100 completions vs MolmoAct2 0/100; mean progress 46 vs 12 out of 100.