Advancing Creative AI: Introducing Veo 2 and the Latest Imagen 3
Earlier this year, we launched Veo, our video generation model, alongside Imagen 3, our cutting-edge image generation model. Since then, creators and businesses alike have embraced these tools to unlock new creative possibilities. YouTube creators have explored video backgrounds for Shorts, enterprise clients have enhanced creative workflows on Vertex AI, and artists have used VideoFX and ImageFX to tell compelling stories. Collaborating with filmmakers, businesses, and artists, we continue to refine and advance these technologies.
Introducing Veo 2: Next-Level Video Generation
Today, we’re excited to unveil Veo 2, our latest video generation model, delivering state-of-the-art results. Available through VideoFX, ImageFX, and our newest Google Labs experiment, Whisk, Veo 2 sets a new standard for video creation.
Veo 2 excels in generating high-quality videos across various subjects and styles. Human raters have confirmed its superior performance compared to leading models. It demonstrates a deeper understanding of real-world physics, human movement, and expressions, producing videos with exceptional detail and realism.
Veo 2 also comprehends cinematic language. Specify a film genre, camera lens, or cinematic effect, and it responds with visually stunning outputs. Whether you request a low-angle tracking shot through a bustling scene or a close-up of a scientist at work, Veo 2 delivers. Mention "18mm lens" for a wide-angle shot or "shallow depth of field" for a focused subject with a blurred background. Its ability to minimize common video-generation issues like extra fingers or unexpected objects results in more believable outputs.
Committed to safety and responsible AI development, we have taken a measured approach in scaling Veo 2’s availability. This allows us to monitor, evaluate, and enhance its quality and safety. All outputs include an invisible SynthID watermark, ensuring they are clearly identified as AI-generated to reduce misinformation and misattribution.
Starting today, Veo 2’s capabilities are expanding within VideoFX, with broader availability planned for YouTube Shorts and other products in the coming year. Join the Google Labs waitlist to experience it first-hand.
Advancing Image Creation with Imagen 3
We’ve also upgraded Imagen 3, which now generates even more visually striking and well-composed images. Its enhanced capabilities include producing diverse art styles — from photorealism to impressionism, abstract art to anime — with improved accuracy, detail, and texture rendering.
In head-to-head comparisons, Imagen 3 achieved top-tier results according to human raters. This enhanced version will begin rolling out globally through ImageFX to over 100 countries starting today. Visit ImageFX on Google Labs to explore its new capabilities.
Introducing Whisk: Creative Remixing Through Images
Our latest experiment, Whisk, merges AI-powered visual understanding with creative image generation. With Whisk, you can upload or create images that reflect your vision’s subject, scene, and style. Then, blend and remix these elements to craft something truly unique — from digital plushies to custom stickers and enamel pins.
Whisk integrates Imagen 3 with Gemini’s advanced visual understanding capabilities. Gemini generates detailed captions for your images, which are then used by Imagen 3 to produce personalized outputs. This seamless combination enables intuitive and playful remixing for limitless creative possibilities.
Explore Veo 2, Imagen 3, and Whisk through Google Labs and see how they can bring your creative ideas to life.
Validate your login
Sign In
Create New Account