Google Launches Veo 2 Video Generation AI Model for Advanced Gemini Users
On Tuesday, Google announced that the Veo 2 artificial intelligence (AI) model would be integrated into its Gemini chatbot. This feature will be available exclusively for paid subscribers of Gemini. With this advanced AI model, eligible users can create short video clips, lasting eight seconds, simply by using natural language text prompts. The Veo 2 model first emerged in December 2024 as a successor to the previous Veo AI model and is also utilized in the Vertex AI platform, enhancing YouTube’s Dream Screen feature.
Creating 720p Videos with Veo 2 in Gemini
In an official blog post, Google provided details about the Veo 2 video generation AI in Gemini. The rollout is global and will support all languages available in Gemini, but access is limited to Gemini Advanced subscribers. Those using the free version of Gemini will not be able to utilize this model.
Users can find Veo 2 in the model selection menu within both Gemini’s web client and its mobile applications. The model generates eight-second video clips based solely on user-defined text prompts. Each video is produced in 720p resolution with a 16:9 aspect ratio and can be easily downloaded in MP4 format. Google has instituted a limit on the number of videos a user can generate each month and will notify users as they approach this limit.
Additionally, Google has incorporated sharing capabilities for users, allowing them to post their generated videos on social media platforms, including TikTok and YouTube, with just one tap on the share button in the Gemini app. Although the new feature is gradually becoming available, it may take several weeks for all eligible users across the globe to gain access.
Key Features of Veo 2
When Veo 2 was launched, Google highlighted several enhancements aimed at improving the quality and realism of the generated videos. These improvements include:
- Increased Realism: The AI is designed to produce visually realistic and detailed video elements.
- Cinematography Understanding: Users can provide specific instructions regarding camera movements, lens selections, and cinematic effects, which the model can accurately interpret and execute.
- Reduced Hallucinations: Google claims that the model has been optimized to hallucinate less, meaning that artifacts like extra fingers, strange limbs, or unexpected objects should occur less frequently in generated videos.
These features indicate Google’s push toward creating more sophisticated and user-friendly AI tools. By integrating Veo 2 into Gemini, they aim to give content creators an innovative way to generate engaging video content quickly and with minimal effort. As the technology evolves, it is likely that we will see even more enhancements that could further revolutionize how users create and share videos online.
