Internet reports that at 1 a.m. today, Google (GOOG.O)DeepMind finally integrated the long-awaited Veo2 into the GeminiApp application and made it fully open for use. Veo2 can generate up to 8 seconds of 720P movie-level video. It is excellent in camera operation, text semantic restoration, physical simulation, action consistency, etc. It also supports the function of converting pictures to video. According to test data released by Google, Veo2 has surpassed Sora, Creole 1.5, MetaMovieGen and Minimax in terms of user preferences and prompt restoration. In addition, starting today, developers can use Veo2 through the API in Google AIStudio. (AIGC Open Community)