Verne Robotics V0.5 open-weight world action model
Robustness to scene changes, not raw benchmark scores, is what separates deployable robot models from lab demos, and V0.5 posts a much smaller drop than π0.5 under distribution shift. Robotics teams evaluating open-weight manipulation models can benchmark V0.5 against π0.5 in their own cluttered environments. Its base on the Wan2.2 video generator also shows video world models doubling as action predictors.
Verne Robotics released V0.5, an open-weight world action model that takes camera input and outputs the next second of arm, gripper and mobile-base motion. The model learned new industrial tasks from under 2 hours of demonstrations and was trained on over 1.7M demonstrations built on the Wan2.2 video generator. On the RoboTwin 2.0 benchmark, randomized clutter, lighting and textures cut V0.5's success rate from 81.0% to 73.4% in Verne's tests, while Physical Intelligence's π0.5 fell from 70.7% to 46.0%. In simulation with unseen clutter and lighting, V0.5 lost 7.6 points versus 24.7 for π0.5.
Brilliant release by @VerneRobotics
V0.5, an open-weight world action model, that looks through cameras and outputs the next second of arm, gripper and mobile-base motion.
The big deal is that it learned new industrial tasks from under 2 hours of demonstrations.
> Robot models usually stumble when a room changes. With clutter and new lighting that neither model trained on, Verne Robotics’ V0.5 lost 7.6 points in simulation, while Physical Intelligence’s π0.5 lost 24.7.
> On RoboTwin 2.0, randomized clutter, lighting and textures cut V0.5 from 81.0% to 73.4% in Verne’s tests, while Physical Intelligence’s π0.5 fell from 70.7% to 46.0%.
> Built on the Wan2.2 video generator, V0.5 learned from over 1.7M demonstrations to predict upcoming camera frames alongside the motion producing them.
https://t.co/BaTwYMcNsj
https://t.co/8WKzZHXJIw
