Runway, a New York City-based startup, has introduced a groundbreaking feature in its latest AI video model, Gen-3 Alpha. This new capability allows users to generate realistic videos from both text prompts and still images within seconds. The announcement marks a significant advancement in the field of generative AI, showcasing the rapid evolution of the technology.
Get started with Gen-3 Alpha Image to Video.
Learn how with today’s Runway Academy. pic.twitter.com/Mbw0eqOjto
— Runway (@runwayml) July 30, 2024
Gen-3 Alpha offers users the option to create video content driven by still images.
Today we are releasing Gen-3 Alpha Image to Video. This update allows you to use any image as the first frame of your video generation, either on its own or with a text prompt for additional guidance.
Image to Video is major update that greatly improves the artistic control and… pic.twitter.com/OieDwMIspz
— Runway (@runwayml) July 29, 2024
By uploading an image to the Runway platform and using text prompts to guide the video generation process, users can create 5 or 10-second videos, priced at 20 and 40 credits respectively, as part of a subscription model. Beta tests have shown impressive results, with users reporting successful video generation in under a minute.
One user shared a brief video they created, highlighting the seamless and fluid execution.
So in the last day Midjourney updated to a new model and Runway allowed you to use images to turn into video clips. Fascinating that I can go from an idea to a 10 second clip in under 2 minutes👇
Of course this is far from perfect or controllable enough for serious use, but cool pic.twitter.com/0t4XIvg8PC
— Ethan Mollick (@emollick) July 30, 2024
Runway has explained that the AI naturally attempts to block video creation from explicit or recognizable images, ensuring a level of security and ethical consideration in its operations. According to Cristóbal Valenzuela, co-founder and CEO of Runway, Gen-3 Alpha significantly improves artistic control for users, allowing them to mold and guide the output more effectively.
With natural language input and image references, creators can explore fascinating visual storytelling avenues that have previously been cumbersome and resource-intensive. “Today we are releasing Gen-3 Alpha Image to Video.
Gen-3 Alpha’s video creation features
This update allows you to use any image as the first frame of your video generation,” said Valenzuela, promoting the versatility that comes with the new iteration. The AI video landscape has become competitive, with several companies vying for supremacy. OpenAI’s Sora, which remains in a closed beta phase, promises longer video capabilities of up to one minute but is not yet publicly available.
In contrast, Runway’s model can generate outputs quickly and is accessible to users, positioning it as a strong contender in the market. However, the cost of generating AI videos raises crucial questions about accessibility and democratization within the field. Runway’s pricing model presents a notable barrier for some potential users, especially those who may wish to create multiple projects.
Director Paul Trillo recounted that creating his music video required over 700 generations at a significant overall cost. As AI tools become more entrenched in the creative landscape, it’s imperative for the industry to consider how to make these technologies equitable. The pricing structure, particularly for graphic designers and content creators who may rely heavily on these tools for revenue, may need reassessment to avoid alienating potential users.
Runway’s Gen-3 Alpha represents a significant milestone within the AI space. With its ability to transform still images into dynamic videos at a quality that rivals traditional filmmaking methods, the product exemplifies how generative AI is reshaping the landscape of visual media. However, as the conversation around ethical AI continues, companies must navigate a path that balances innovation with respectful practices.
