Runway unveiled the "Gen 4.5" model
Runway, an AI startup, unveiled on Monday its new video creation AI model, Gen 4.5, which outperforms similar models from OpenAI and Google in an independent test.
Gen 4.5 allows users to create high-definition videos based on text commands describing desired movements and actions.
Runway explained that the model demonstrates a strong understanding of physics, human movement, camera movement, and cause-and-effect relationships, according to a CNBC report reviewed by MobiAlo.
The model topped the Video Arena rankings, run by Artificial Analysis, a company specializing in AI model analysis and comparisons. To determine the rankings of text-to-video models, users compare two different models and vote for the best one without knowing the companies behind them.
Google's Veo 3 model ranked second on the list, while OpenAI's Sora 2 Pro ranked third. Runway AI ranked seventh.
“We were able to outperform trillion-dollar companies with a team of 100 people,” Runway CEO Cristóbal Valenzuela told CNBC in an interview. “You can reach advanced levels as long as you are focused and extremely hardworking.”
Founded in 2018, Runway researches and develops AI models for video and other applications. These models are trained on video and observational data to better reflect how the physical world works.
The company’s clients include media organizations, studios, brands, designers, creators, and students. Its market capitalization has reached $3.55 billion, according to PitchBook.
Gen 4.5 will be rolled out gradually but will be available to all Runway clients by the end of the week. Valenzuela said it is the first of several major releases the company will make.
Source: Reuters
