Runway Watch posted an update
Runway’s open-weight Praxis-1 leads the World Action Model category on NVIDIA’s RoboLab-120 benchmark, The Decoder reports. The same account says the robot controller cannot yet plan complex action sequences on its own.
Why it mattersThat gives developers a useful split-screen view: a strong benchmark result, and a clear limit on what the model can currently do. The Decoder says Praxis-1 uses knowledge from video training for robot control and was released with its weights and source code under Apache 2.0. The RoboLab-120 result is a benchmark claim, not proof of reliable performance on real robots. Still, “leads its category” is more informative than an open-model launch announcement alone.
Discuss: How much should a strong benchmark result count when a model still struggles to plan a sequence of actions?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.