Solaris
Solaris is a real time interactive world model that generates video frame by frame in response to user input. It takes text prompts and images as conditioning and treats interactions such as clicks, drags and typing as conditioning for the next frame, so an on screen interface responds continuously as it is used. It renders at 720p and is built on the Gen 4.5 video model, using autoregressive frame generation with few denoising steps to stay within interactive latency.
- Type
- Video
- Weights
- Closed
More from Runway
9
-
Act Two
Video
Act-Two is a narrative-aware video generator that turns scene directions into cohesive sequences. It maintain…
-
17k DL
Aleph
Video
Aleph is a unified creative foundation model that spans text, image, and video generation. It emphasizes cons…
-
Aleph 2
Multimodal
Aleph 2.0 is Runway’s closed-source in-context video editing model for modifying existing videos while preser…
-
Gen 1
Video
AI model by Runway
-
Gen 3 Alpha
Video
Gen-3 Alpha is Runway’s next-gen video model for text/image-to-video and video editing. It delivers sharper m…
-
Gen 4
Video
Gen-4 is a next-generation video model that raises fidelity, temporal stability, and physical realism while e…
-
Gen 4 Turbo
Video
Gen-4 Turbo is the high-throughput variant of Gen-4, trading a touch of peak fidelity for much faster iterati…
-
Gen 4.5
Video
Runway Gen-4.5 is Runway's latest text-to-video and image-to-video model, ranked #1 on independent benchmarks…
About Runway
1
Runway provides generative AI tools for creating and editing video, images, and audio.