Robots are starting to perform new tasks without task-specific robot demonstrations.
LiberAI’s Liber-0 achieved 36.0% overall success across 60 GPT-6-proposed tasks and 600 trials. The evaluated checkpoint was not trained on robot demos collected specifically for those tasks.
That included 10.5% success on challenging multi-step manipulation.
The bigger question: how do we turn human experience into robot skill?
Human video shows what people do, but doesn’t directly specify commands a robot can execute. Our bodies see and move differently.
Liber-0 addresses that gap through:
Action grounding. Glove recordings capture hand poses and trajectories, connecting video with measured motion.
Embodiment alignment. Human-like robot movement and body-aware action prediction help adapt those skills to robot execution.
Pretrained on 2.2M+ hours of human experience, Liber-0 points toward a way to build robot capabilities with fewer task-specific robot demonstrations.
Paid partnership with @liberai_robot
https://t.co/sicKKuBIiC
We’re renaming our Max plans:
Max 5x → Max
Max 20x → Ultra
Max is still $100/month, and Ultra is still $200. If you’re a subscriber, there’s nothing you need to do.
We’re also removing the 5-hour limit for verified Ultra accounts, starting today with most existing subscribers. Newer accounts keep a temporary limit until they’re verified, which helps us prevent fraud and abuse. You'll get an email when it's removed.
Weekly limits remain unchanged. View our plans here: https://t.co/niRpyGF5Me