XAI releases Grok Imagine Video 1.5: supports simultaneous audio and video generation, doubling speed According to Beating monitoring, xAI has officially released the image and text generation video model Grok Imagine Video 1.5, which is fully available on the API (grok-imagine-video-1.5), web ((grok. com)/imagine), and mobile client. The model achieves integrated synchronous generation of audio and video, synchronously generating sound effects, environmental sounds, and character dialogues in a single inference stage, improving speech clarity and optimizing lip sync. At the same time, the model improves the physics engine and motion consistency, enhances the credibility of object motion and physical weight over long periods of the lens, and reduces artifacts such as image distortion. In terms of generation speed, the lightweight version of Video 1.5 Fast reduces the time required to generate a 6-second 720p video to about 25 seconds. Web based workflow synchronization update: Added project management (Projects) to classify and organize materials, supports multiple agents to run multiple prompt words in parallel, and provides media library semantic search (Search). The team of digital artist David Thompson used Grok Imagine 1.5 to create a fully AI generated movie trailer called 'Odyssey'. [Original link]