Praxis-1

Overview

Runway introduced Praxis-1 on September 30, 2026, its first open-weight world-action model, built on the large-scale video pretraining behind Runway's general world models and adapted for robot control. A single policy runs across different robot embodiments, including bimanual rigs, six-axis arms, and mobile bases, with only light fine-tuning, on tasks ranging from simple pick-and-place to manipulating deformable objects like packing gift bags. Early partners Noble Machines, Standard Bots, and Ultra are running Praxis-1 on their own hardware via early access through Runway's robotics team; Runway says public weights will follow in the coming months.

This is a sourced directory entry. Features and capabilities are reported in the linked sources; this entry does not document independent Aymorphic testing.

Key features

• Open-weight world-action model trained primarily on video data
• One policy generalizes across bimanual rigs, 6-DoF arms, and mobile bases without retraining
• Handles both rigid pick-and-place and deformable-object manipulation
• Early access live with hardware partners Noble Machines, Standard Bots, and Ultra

Use cases

• Cross-embodiment robot manipulation policies without per-robot retraining
• Bootstrapping robot control from video pretraining where robot demonstration data is scarce
• Hardware vendors adding general-purpose manipulation to existing robot platforms

Pricing

Early access only via Runway's robotics team; public Apache-style open weights promised in the coming months, pricing not yet disclosed.

Pricing reflects the source record and may change. Confirm current terms with the provider.

Dates and sources

Released

30 Sept 2026

Record updated

10 Oct 2026

Record last checked

10 Oct 2026

• runway.com/research/introducing-praxis-1
• therobotreport.com/runway-introduces-praxis-1-world-action-model-robotics/
• roboticsandautomationnews.com/2026/10/01/runway-moves-into-robotics-with-open-weight-praxis-1-ai-model/105402/