What Runway announced: Gen 4.5 video model
According to a report by CNBC, Runway today released its new video-generation model Gen 4.5. The company claims that the model can generate high-definition video from text prompts, demonstrating a strong understanding of motion, human actions, camera choreography, and cause-and-effect, effectively simulating realistic physical behavior.
Runway claims that Gen 4.5 has significantly improved its grasp of real-world physics and video consistency, enabling more believable, cinematic outputs than previous models.
🏆 Independent Benchmark — Video Arena
According to coverage of independent evaluations, Gen 4.5 topped the “Video Arena” benchmark. In this blind-review ranking, users compare video outputs from different models and vote for the one they consider better. Gen 4.5 reportedly beat models from major players such as Google (Veo 3) and OpenAI (Sora 2 Pro).
In that ranking, Google’s Veo 3 placed second, while OpenAI’s Sora 2 Pro ranked seventh.
🗣 What Runway’s CEO said
Runway’s CEO Cristóbal Valenzuela told CNBC that their success shows even a small company of ~100 people can outperform multi-trillion-dollar tech giants — if focused and diligent. He said the codename for Gen 4.5 is “David,” referencing the “David vs. Goliath” idea.
Valenzuela argued that in the current era, where efficiency and innovation matter, it’s essential that generative AI not become monopolized by just two or three companies.
🧑💻 What is Runway, and Why This Matters
- Runway is an AI startup founded in 2018, focusing on video generation, “world models,” and related AI research.
- Its customers include media agencies, film studios, brands, designers, creators, and students.
• • The release of Gen 4.5 could mark a significant step forward for text-to-video AI, improving realism, motion accuracy, and cinematic quality, which may make AI-generated videos more usable for creators, studios, or marketing teams.
The above content is compiled by ModeZone, a fashion and entertainment magazine.
