Adobe Firefly AI to Get Video Capabilities in 2024 — Still No Timeline for OpenAI Sora
OpenAI appears to be in no rush to release Sora but may lose its competitive edge if it waits too long. | Credit: AaronP/Bauer-Griffin / Getty Images.
Key Takeaways
- Adobe has announced its Firefly Video Model.
- The AI video generator will be launched in Premiere Pro later this year.
- As the text-to-video market heats up, OpenAI has still not committed to a release date for Sora.
Eight months after its initial reveal, OpenAI still hasn’t committed to a release date for its AI video generator Sora. In the meantime, rival platforms like Runway and Luma have filled the space.
Meanwhile, Adobe has announced plans to launch Firefly Video Model—an AI video generator that will be beta-launched as a feature on Premiere Pro later this year.
Adobe Expands Firefly With Text-To-Video Functionality
Adobe first launched its computer vision model Firefly in March 2023, and since then, it has been integrated across the company’s software suite, powering features like Generative Fill in Photoshop, Generative Remove in Lightroom, Generative Shape Fill in Illustrator, and Text-to-Template in Express.
The latest model builds on that legacy, introducing video-generating capabilities to Premiere Pro. Given Adobe’s history of expanding Firefly across its software suite, programs like After Effects and Premiere Rush could also be enhanced with the new model further down the line.
Adobe said Firefly Video would be available in beta later this year without committing to a specific date.
Meanwhile, there is still no update on the release timeline for OpenAI’s Sora.
No Launch Date for OpenAI Sora
OpenAI initially teased its generative video model Sora in February 2024, attracting significant media and public attention with a series of high-fidelity AI videos.
While the platform has been made available to a limited number of artists and creators, ongoing concerns around its potential for abuse and extensive safety testing have held up a wider public release.
In a comment initially shared with The Wall Street Journal, OpenAI CTO Mira Muratti said the company will “definitely” release the model this year. However, she added that “we won’t be releasing anything when we won’t be confident on how it might affect global elections,” leading to speculation that the firm will wait until after the U.S. elections in November.
The Competitive Landscape
Granted, OpenAI may have legitimate reasons not to rush out Sora before it is ready, but it risks losing a competitive edge if it waits too long.
When OpenAI first debuted Sora’s video-generating capabilities, they were much more impressive than the outputs from existing solutions on the market. But other AI developers have since caught up.
Alongside the impressive Firefly outputs recently shared by Adobe, newer models from Runway and Luma have made significant leaps in video quality, generating more detailed, smoother clips that rival the ones created by OpenAI.
Meanwhile, a Chinese startup launched Kling in July 2024, approaching Sora-level visual realism for videos of up to 10 seconds.
Adobe could also have an important advantage over OpenAI. Whereas Sora has been dogged by questions over the sources of its training data, Adobe boasted that the Firefly Video Model “is designed to be commercially safe and is only trained on content we have permission to use.”
OpenAI CTO’s Evasive Response Raises Questions About Sora’s Data Sources
Sora Release Date: OpenAI’s Video Generator Public Timeline, Details
