OpenAI и Google усовершенствовали свои видеомодели, добавив новые функции и улучшения Translation: OpenAI and Google Enhance Their Video Models with New Features and Improvements

The startup OpenAI and tech giant Google have unveiled updated versions of their AI models for video generation.

In the web version of Sora 2, Pro subscribers now have access to storyboards. Everyone can create videos of up to 15 seconds in the app and on the site, while premium subscribers can produce videos of up to 25 seconds online.

Google has introduced Veo 3.1, which features enhanced sound, more precise editing tools, and improved video generation quality from images. The neural network retains the functionality of Veo 3, but it offers a more realistic visual output and responds more accurately to prompts. This version has been integrated into the Flow video editor.

Users can now seamlessly incorporate new objects into their videos. Soon, there will be an option to remove existing items directly within Flow.

The previous version of Veo 3 already included features for adding reference images to control characters, generating video from the first and last frames, and continuing an existing video based on the latest frames. The new release adds sound to all these capabilities, making clips feel more «alive.»

It’s worth noting that in September, Google launched an affordable plan called Google AI Plus in over 40 countries, including Ukraine, Moldova, Kyrgyzstan, Uzbekistan, Egypt, and others.