xAI Grok Imagine 1.5 Preview Release

xAI has released grok-imagine-video-1.5-preview, an image-to-video model available via the xAI API. This model allows users to transform a single still image into cinematic video by combining a starting frame with natural-language prompts describing the motion, atmosphere, and physics.

Image-to-Video Generation Capabilities

grok-imagine-video-1.5-preview generates fluid, cinematic video clips at resolutions up to 720p. The model is designed to maintain strict fidelity to the source image, ensuring that lighting and detail from the input frame are preserved rather than reinterpreted during the animation process.

Motion and Camera Control

Users can direct the video generation process using natural-language prompts to specify:

• Camera Movement: Precise control over how the camera moves through the scene.
• Pacing: The speed and timing of the animation.
• Soud Design: Descriptions of the accompanying audio elements.

Sequence and Consistency

The model supports the creation of longer scenes by chaining together animated frames. By staging each frame and animating it sequentially, users can maintain a consistent visual look across an entire project.

API Access and Implementation

grok-imagine-video-1.5-preview is currently available in preview via the xAI API, allowing developers to integrate image-to-video animation capabilities into their applications using a few lines of Python code.