Summary
- Google’s new Veo 3 video generator is generating significant excitement, seen by some as a potential leap ahead of OpenAI in AI video, thanks to its integration of Sora-like video with built-in audio and accurate lip-syncing.
- A key advancement in Veo 3 is the ability to generate dialogue directly within the text prompt and maintain consistent characters across multiple video segments, simplifying workflows and enhancing creative possibilities.
- Access to Veo 3 is currently exclusive to Google’s AI Ultra plan, priced at $250 per month, which also includes other premium AI features and storage.
Google I/O’s standout announcement, without a doubt, was Android XR and the tech giant’s early demonstrations of the hardware running it. It intertwines with the general theme seen across the developer conference this year — AI, and its real-world use cases.
Android XR glasses are poised to use Gemini for applications like real-time translations, directions, scheduling entries, and more, but that’s only one facet of Google’s AI announcements. The tech giant also showed off AI-powered enhancements for virtual try-on, agentic checkout for AI Mode, upgraded AI models, and most importantly, a new version of its Veo video generator.
The tech giant’s new “state-of-the-art” video generation model has made a buzz, but not in the truest sense. That can largely be attributed to AI fatigue, but beneath the noise, Veo 3 is actually a significant leap forward. It might be the first time that Google has a better AI product than OpenAI. Most AI subreddits, which are often skeptical of Google’s AI endeavors, seem to agree, and it might just be time to pay attention.
At its core, Veo 3 is essentially a fusion of Sora’s video generation prowess, paired with audio generation and accurate lip-syncing capabilities, and users are already starting to experiment with it for personal projects.
There’s a reason why access to Veo 3 is locked behind a new $250 per month plan
User MetaPuppet shared a two-and-a-half minute project that they made using Veo 3, and it is nothing short of impressive. For reference, Veo 3 lets you generate eight-second-long clips, which means the user was able to generate multiple small clips and piece them together in a cohesive manner, with consistent audio, and most importantly, character continuity.
Forget fidelity and physics for a second. The real game-changer? Being able to generate dialogue right in the text prompt. What used to take two extra steps now happens instantly — and the quality? Unreal.
Another user created a minute-long SWAT team operation, with frame-perfect lip-syncing, highlighting just how far we’ve come from the old ‘Will Smith eating spaghetti‘ days, while further blurring the lines between real and artificially generated video content.
For what it’s worth, access to the new model isn’t cheap. Veo 3 is locked behind Google’s new AI Ultra plan, which costs $250 a month, but also unlocks access to other top-end models, premium features, 30 TB of storage and access to YouTube Premium.
