Breaking News! Runway's Gen-3 Now Open to All, Rivals Sora!

Creati.ai AI News - July 2, 2024

In the early hours of July 2, the renowned generative AI platform Runway announced on its official website that its text-to-video model Gen-3 Alpha is now available to all users. The speed of this rollout is astounding, coming just a day after its initial limited release.

Using Gen-3 is incredibly simple: users input a text description, such as "Enter a jungle, discover a strange multidimensional portal, sci-fi effect, first-person view. Open the door, find a cute little white rabbit," and the model generates a video accordingly. Currently, Gen-3 can produce 11-second 720P videos without background music, though future updates may add this feature. For now, it’s perfect for creating microfilms and promotional ads, with users able to add music using tools like Suno or Uido.

Here’s a video made by a user using multiple Gen-3 generated clips to create a food introduction video. The saturation, lighting effects, motion consistency, continuity, and colors are all outstanding. This is due to Gen-3's physical simulation capabilities, which ensure that the generated content adheres strictly to real-world characteristics.

Examples of videos made with Gen-3 include:

• A promotional video for a microfilm about "Trash Monsters," which boasts nearly professional-level storytelling, camera language, and structure.

• A creative short film about humans and nature, featuring smooth camera movements and scene transitions.

• A group of happy friends playing board games, complete with dynamic fight scenes and impressive flame effects.

• A touching story about the loneliness of a black bear, where the character eats, sleeps, and travels alone, though the bear’s appearance changes several times.

• A heartwarming animal short film depicting the search for tranquility amidst busy lives, emphasizing the joy of simple pleasures and inner peace found in the taste of food.

The rapid availability of Gen-3 has excited many users, with some claiming it has already surpassed Sora. It's worth noting that Sora has been in invitation-only testing for over four months, with no public release in sight. Many now consider Gen-3 the best text-to-video model, reinstating Runway as the leader in this domain after Sora’s brief reign.

Runway’s Creative Director expressed immense excitement, stating that Runway has made history and is set to revolutionize the text-to-video space once again. However, unlike previous versions and Luma, Gen-3 does not offer a free trial, likely due to computational constraints, with a minimum usage fee of $12 per month. This decision might have been influenced by Luma’s server crash on its first day of free access.

Many users are unhappy with this paywall, criticizing Runway for not providing even a limited free trial. They argue that a small number of free uses per day would be better than jumping straight to paid access.

Despite this, Runway's Gen-2 and other products remain free, and the company has promised that Gen-3 will include advanced features such as motion brushes, camera controls, and director mode to help users create even more polished videos.

This development marks an exciting moment in the text-to-video model arena, with Runway’s Gen-3 poised to set new standards and inspire further innovation.