xAI releases Grok Imagine Video 1.5: Supports synchronized audio and video generation, doubling the speed.

According to Beating, xAI has officially released Grok Imagine Video 1.5, an image and text generation video model, which is now fully available on the API (grok-imagine-video-1.5), web platform (grok.com/imagine), and mobile client. The model achieves integrated audio and video generation, simultaneously generating sound effects, ambient sounds, and character dialogue during a single inference phase, improving speech clarity and optimizing lip-sync. Simultaneously, the model improves the physics engine and motion consistency, enhancing the credibility of object movement and physical weight over long shot periods and reducing artifacts such as image distortion. In terms of generation speed, the lightweight version, Video 1.5 Fast, generates a 6-second 720p video in approximately 25 seconds. The web platform's workflow has also been updated: a new Projects feature has been added to categorize and organize materials, supporting multiple agents running multiple prompts in parallel, and providing semantic search for the media library. Digital artist David Thompson's team used Grok Imagine 1.5 to create a movie trailer for "Odyssey" that was entirely generated by AI.