Archive · 12 Dec 2025
Runway Launches First World Model
Runway, a tech company, has launched its first world model called GWM-1.
This new model is designed to help AI understand and simulate the real world, making it useful for things like robotics and gaming.
Runway also upgraded its Gen 4.5 video model to include native audio, character-consistent long-form generation, and multi-shot storytelling.
The GWM-1 model has three specialized versions: GWM-Worlds for creating interactive spaces, GWM-Robotics for training robots, and GWM-Avatars for creating lifelike digital humans.
These tools are meant to be used both for research and in real-world industries.
Runway's latest releases show that world models and video systems are moving from theory into practical use.
Runway launched its first world model, GWM-1, alongside upgrading Gen 4.5 with native audio.
GWM-1 focuses on frame-by-frame prediction for a sense of physics, movement, and environmental continuity.
Three specialized versions of GWM-1 were unveiled: GWM-Worlds, GWM-Robotics, and GWM-Avatars.
Gen 4.5 now supports native audio, character-consistent long-form generation, and multi-shot storytelling.
Runway aims to serve both research and real-world industries with its latest releases.
- Who
- Runway, a tech company
- What
- Launched its first world model, GWM-1, and upgraded Gen 4.5 with native audio
- When
- Recently, with Gen 4.5 debuting earlier this month
- Where
- Not specified, but likely based in the USA
- Why
- To position itself at the front of the industry shift towards deployable, production-grade tools
- Company
- Runway
- New Model
- GWM-1
- Model Versions
- GWM-Worlds, GWM-Robotics, GWM-Avatars
- Gen 4.5 Upgrades
- Native audio, character-consistent long-form generation, multi-shot storytelling
- Industries Targeted
- Robotics, life sciences, gaming, virtual production
- Availability
- GWM-1 suite for research and industries, Gen 4.5 for paid-plan users