New paper tests GPT-6-Astra on 12 robotics benchmarks
A new paper tests GPT-6-Astra on 12 robotics benchmarks; it approaches saturation on the LIBERO suite, reaching 81.9% on RoboDojo and 87.3% on DexJoCo. However, a single LIBERO-Ultra-Long episode consumes about 9M tokens and takes nearly 2 hours, and on real robots, more steps cannot be traded for success. The authors argue that general-purpose models can currently serve as robotics controllers for research, but task-specific demonstrations, high compute demands, and physical limitations still hinder deployment.
Why read this
AI score 58
The paper tests a general-purpose model on 12 robotics benchmarks, simultaneously reporting high scores along with the cost and physical limitations on real robots.
Topics
Body
New paper finds that GPT-6-Astra nearly saturates easy robot benchmarks but still fails to verify physical success and recover from irreversible mistakes. And at a cost too high for real use.
Robot manipulation and navigation have been dominated by task-specific models. The question was whether 1 general model could handle them all without weight updates.
The authors tested GPT-6-Astra on 12 robot benchmarks. It studied training demos into a frozen knowledge package and got up to 10 times the usual action steps to recover.
It neared saturation on the LIBERO suite and reached 81.9% on RoboDojo and 87.3% on DexJoCo. But a single LIBERO-Ultra-Long episode used about 9M tokens and nearly 2 hours, and on real robots it could not trade more steps for success.
Treat general models as capable robot controllers for research, but expect task-specific demos, heavy compute, and physical limits to block deployment for now.
– arxiv. org/abs/2610.22471
Title: "Has GPT-6-Astra Achieved Embodied Intelligence?"