A mug on a table. A tray beside it. Predicting the picture of a successful move is one problem. Producing the robot actions that get there is another. This short visual explainer introduces Runway Praxis-1, a world action model that Runway describes as predicting video and actions together. Its open-weight release arrived on November 11, 2026, with weights and inference/fine-tuning code under Apache 2.0. We follow one original animated example through reach, grip, lift, move and release. The engineering-bench animation is our illustration of the concept. It is not footage generated by Praxis-1, a benchmark, a hardware demonstration or evidence that we ran the model. Runway reports partner demonstrations on different robots. Adapting a model to a particular embodiment, camera setup and task still matters; an open release does not mean every robot will work out of the box. We deliberately keep company claims separate from our own animation and do not repeat leaderboard superlatives. Official sources: https://runway.com/news/company-news/praxis-1-is-now-available https://runway.com/research/introducing-praxis-1 https://github.com/runwayml/praxis-1-oss Which part looks hardest to you: seeing the object, planning the move, keeping contact, or recovering when reality differs from the prediction? Original authored animation, synthetic English narration, and original music/sound design. Produced in Framewright. No affiliation with or endorsement by Runway is implied. More visual stories: https://www.youtube.com/@ABV_Creative TikTok: https://www.tiktok.com/@lumierenoire.films #Robotics #Praxis1 #Runway