Runway released the powerful Gen-4 model for video generation
Runway has released the Gen-4 model, which is able to generate videos with one of the highest levels of visual quality among competitors. It is already available for corporate clients and individual users of the Runway platform.
TechCrunch writes about it.
What the new model from Runway can do
The new Gen-4 model can generate stable images of characters, locations, and objects in different scenes without violating the integrity of the environment and reproducing individual elements from different positions and angles. Users can provide image models of a particular character and make a description of the shot or composition they would like to get.
"Gen-4 can utilize visual references, combined with instructions, to create new images and videos utilizing consistent styles, subjects, locations, and more, [a]ll without the need for fine-tuning or additional training," Runway claims in its blog.
Today we're introducing Gen-4, our new series of state-of-the-art AI models for media generation and world consistency. Gen-4 is a significant step forward for fidelity, dynamic motion and controllability in generative media.
— Runway (@runwayml) March 31, 2025
Gen-4 Image-to-Video is rolling out today to all paid… pic.twitter.com/VKnY5pWC8X
The Gen-4 model can maintain the correspondence of characters in different lighting conditions, as well as simulate realistic physics in the frame. The developers of the model said that they trained the new tool on a significant amount of video material without disclosing the source of the data in order not to lose competitive advantage.
Gen-4 sets a new standard for video generation and is a marked improvement over Gen-3 Alpha. It excels in its ability to generate highly dynamic videos with realistic motion as well as subject, object and style consistency with superior prompt adherence and best-in-class world… pic.twitter.com/w9ACO5boJ7
— Runway (@runwayml) March 31, 2025
However, this approach of the company has already caused lawsuits. The class action lawsuit has already been filed against several companies, including Runway, for using copyrighted images without permission to train algorithms.
The startup itself is backed by investments from Salesforce, Nvidia, and Google and mainly specializes in video tools. The company's competitors are the giants OpenAI and Google with their Sora and Veo 2 models, respectively. Runway aims to maintain its position in the market, which is why it has already signed an agreement with a major Hollywood studio and invested heavily in the creation of films using AI.
As a reminder, ByteDance, which owns TikTok, has introduced the OmniHuman AI video generator. It can create realistic videos based on a single photo.
We also wrote that Google revealed the cost of using its new video generation tool Veo 2. Among all analogs, it is the most expensive to use.