Google&s Veo 2 is currently the best video AI, but it wasn&t publicly accessible until recently. After Google integrated it a few days ago into the YouTube app (for free!) to create YouTube Shorts, Veo 2 is now available for general use.
However, Google is making money from Veo 2&s use: at a price of 50 US cents per second or per minute of generated video, Veo 2 is the most expensive of all current video models. Veo 2 is not directly accessible in Europe, but it can still be used here via meta-services like Freepik or FAL, which offer access to various image and video AI models such as MiniMax, Hunyuan, and Kling 1.6 via API. Generated clips can be between 5 and 8 seconds long (at least with FAL) at a resolution of 720p, an aspect ratio of either 16:9 or 9:16, and 24 frames per second. With FAL, 1 minute of video costs only .
Video AI Prices Compared
For comparison: OpenAI&s Sora (Turbo) costs for 50 five-second videos (250 seconds or about 4 minutes) in the ChatGPT Plus plan, Kling 1.6 (in the monthly standard plan) 660 credits for about , which corresponds to a little over one minute of video in high-quality pro mode or 4 minutes in standard mode. The price, however, is relative if fewer attempts are needed to achieve the desired result because the video AI correctly interprets the prompt and does not produce image errors.
Google Veo 2
And that&s exactly what Veo 2 is supposed to do, according to Google and user reports: through a high understanding of physical relationships and the subtleties of human movement, videos with high detail and realism are generated. Veo 2 also understands cinematic terms and can reliably produce various genres, lenses, or special settings such as a tracking shot from a frog perspective or a zooming close-up of a face via prompt.
Here are some example clips of Veo 2:
;)
