Robots Atlas>ROBOTS ATLAS
Runway Gen-4.5
AI Modelsโ€บRunway Gen

Runway Gen-4.5

Gen-4.5ย ยทย Family: Runway Gen
Runway's flagship video generation model (December 2025): text-to-video and image-to-video, 2-10s clips, cinematic photorealism, advanced motion and physics; #1 on Artificial Analysis Text-to-Video.
โœ“ Activeโœ“ Public accessVideo generation๐Ÿ“ Runway Gen
Release date
1 December 2025
๐ŸขRunwayProducer
Access:APIHostedDeployment:โ˜ Cloud

Overview

Runway Gen-4.5 is Runway's flagship video generation model, introduced on December 1, 2025 as the four-and-a-half generation of the Gen line. It supports text-to-video and image-to-video generation, producing clips from 2 to 10 seconds long.

The model stands out for cinematic, photorealistic imagery, precise prompt adherence, advanced motion quality and visual consistency. Runway highlights strong understanding of physics, human motion, camera movements and cause and effect. Gen-4.5 ranked #1 on the Artificial Analysis Text-to-Video benchmark (1247 Elo), surpassing Google and OpenAI models.

The model is available in the Runway app and via the Runway Dev API (credit-based billing). It is also one of the flagship first-party models served by Runway Media Router. Weights are not released (closed model, cloud access).

Classification
Video generation
Family: Runway Gen
Applications
AI video productionCreative advertisingSocial media contentContent generationPersonal visual expression
Access & deployment
APIHosted
Cloud
Weights: Closed
Key parameters
๐Ÿ“ฅ Input: text, image

Technical specification

License
Wlasnosciowa (zamkniete wagi)
Hardware requirements
Cloud-only access (Runway app / Runway Dev API); no downloadable weights. 2-10s clips.
Modalities
โฌ‡ Input
textimage
โฌ† Output
video

Capabilities and applications

Native model capabilities
Video generation
The model's ability to generate video clips from a text prompt, image or another video, with control over length, resolution and visual characteristics.
Category: video
Text-to-video generation
Generating video sequences from a text prompt (and optionally an image), with coherent scene dynamics.
Category: video
Image-to-video
The model's ability to animate a static input image โ€” extending it in time into a consistent video clip according to a description of motion or action.
Category: video
Application domains
AI video productionCreative advertisingSocial media contentContent generationPersonal visual expression

Benchmark results

1 benchmark
Artificial Analysis - Text-to-Video
Elo (ranking) ยท Gen-4.5 ranked #1 on the Artificial Analysis Text-to-Video benchmark with 1247 Elo, surpassing Google and OpenAI models.
#1 (1247 Elo)Elo
๐Ÿ“… 1 Dec 2025
Source: launch coverage (CNBC, 2025-12-01).

Pricing

Technical architecture

Core Architecture
DMDiffusion Model

Deployment and security

๐Ÿ”’ Security / Enterprise
โœ“ Verified enterprise information

Closed model, available via the enterprise-grade Runway Dev platform; access control and governance at platform level (incl. Media Router allow/deny lists).

Updated: 31 Jul 2026โ†— Security documentation

Sources and related pages