Startup Ai Runway on Monday released What he claims is one of the highest quality AI -powered video generators.
The model called Gen-4 is implemented with clients of firms and enterprises. The leak claims that it will probably generate consistent signs, locations and objects in scenes, maintain “coherent world environments” and regenerates elements from various perspectives and positions in the scenes.
“Gen-4 can use visual references, in combination with instructions, creating new images and films using coherent styles, items, locations and others” He wrote in a post on the blog“[a]LL without tuning or additional training. “
GEN-4 is a new standard of video generation and is a significant improvement in comparison with Gen-3 Alpha. It is distinguished by its ability to generate highly dynamic movies with realistic movement, in addition to the subject, object and style with excellent quick observance and the best world … pic.twitter.com/w9aco5Boj7
– Runway (@runwayml) March 31, 2025
Runway, which is supported by investors, including Salesforce, Google and Nvidia, offers a package of video tools AI, including video-generating models equivalent to Gen-4. He faces rigid competition in the space of video generation, including OpenAI and Google. But the company fought for a distinction, concluding a contract with the predominant studio from Hollywood and saving thousands and thousands of dollars to finance videos using a video generated by AI.
Runway says that Gen-4 allows users to generate coherent signs in lighting conditions using a picture of references to those characters. To create a stage, users can provide images of objects and describe the composition of the shot that they wish to generate.
By using visual references in combination with instructions, GEN-4 permits you to create new paintings and movies about coherent styles, objects, locations and others. Allowing continuity and control in your stories.
To test the narrative capabilities of the model, we collected … pic.twitter.com/iyz2baew2u
– Runway (@runwayml) March 31, 2025
“Gen-4 is distinguished by its ability to generate highly dynamic films with realistic movement, as well as the consequence of the object, objects and style with excellent quick observance and the best in the world in the world,” says the company in your blog post. “Gen-4 on the catwalk [also] It is a significant milestone in visual abilities of generative models to simulate physics in the real world. “
Gen-4, like all video generating models, has been trained in a huge variety of examples of flicks to “learn” patterns in these movies to generate synthetic materials. The runway refuses to find out where the training data comes from, partly for fear of dedication to a competitive advantage. But the details of the training are also a potential source of IP -related processes.
An example of the runway is targeted at the suit Imported by artists against IT and other generative firms, which accuse the accused of coaching their models in the field of copyright protected works without permission. The runway claims that Doctrine referred to as permitted use Tarts before legal repercussions. It is not clear yet whether the company will win.
The rates are a bit high for the runway, which is said to lift New funding round It values $ 4 billion. According to the information, the catwalk hopes to hit $ 300 million annual revenues This 12 months after launching products equivalent to API for video generating models.
However, a lawsuit against the catwalk, generative AI video tools threaten to extend the film and television industry as we know. 2024 test Associated by the animation guild, a union representing Hollywood animators and cartoonists, stated that 75% of film production firms that adopted artificial intelligence, reduced, consolidated or eliminated work after the technology was turned on. The study also estimates that by 2026, over 100,000 American entertainment jobs will likely be disturbed by generative artificial intelligence.