Google has unveiled its upgraded Veo 2 video generator model, designed to compete with OpenAI's Sora. The new model promises realistic motion and high-quality video output up to 4K resolution, claiming superiority over other AI video generation platforms.
Short video clips demonstrated by Google showcase Veo 2's ability to generate hyper-realistic videos of animals, food, and animated humans. 'Veo 2 outperforms other leading video generation models, based on human evaluations of its performance,' Google stated, likely referencing competitors like OpenAI's Sora.
Despite impressive visuals, Google acknowledged challenges in maintaining consistency in complex scenes, noting that some motion details may be inaccurate. 'While Veo 2 demonstrates incredible progress, creating realistic, dynamic, or intricate videos remains a challenge,' said Google DeepMind.
Additionally, Google introduced the Imagen 3 model, which produces brighter and more realistic images with enhanced color balance and fidelity. The new Whisk AI model allows users to create images from multiple visual prompts, enabling a creative blend of subjects and styles.