Google's Omni 1.2 Flash tops two video charts: up to 15 seconds in length, generation cost reduced by 30%.

Beating AI News: Google has launched its Gemini Omni 1.2 Flash video generation model. The new version can produce videos up to 15 seconds long, a 50% increase over its predecessor. The cost of 720p video generation has dropped by 30% to $0.07 per second, meaning a 15-second video costs just $1.05 to generate. A key upgrade of Omni 1.2 Flash is its enhanced ability to maintain consistency of characters and visuals during sequential video edits. For example, users can use chat prompts to sequentially change the background of the same video to a forest, stable, and castle, while retaining the original characters, actions, and camera angles. The new version also fixes lip-sync issues for multi-person dialogues and the problem of abrupt visual jumps during continuous camera movements. In the preliminary Arena rankings released on November 5, Omni 1.2 Flash ranked first in both text-to-video and image-to-video categories, and third in video editing. The new model is now available via the Gemini API and integrated with Flow and Google Vids. The 15-second video generation feature of the Gemini App is gradually being rolled out to Google AI Ultra users.