Image and voice references are coming to Grok Imagine Video 1.5

Grok Imagine Video 1.5 combines seven reference images and a voiceprint to stabilize characters. A key asset for shot consistency.

The new version of Imagine Video 1.5 adds generation from a simple text prompt and now supports native 1080p in both text-to-video and image-to-video. Both features are accessible on the web as well as in the iOS and Android apps.

The model can also receive up to seven reference images in a single generation. Each source can be used to maintain a character, a product, or a setting, and then modify the other elements of the scene or the requested action.

A voice reference can be associated with a character's image to preserve their appearance and voice across multiple sequences. This feature starts with the SuperGrok Heavy and SuperGrok Plus subscriptions in the United States, before a gradual rollout to other plans.

The `grok-imagine-video-1.5` API already supports text-to-video, image references, and 1080p rendering. Access to voice references in the API remains subject to a request to xAI.

Imagine Video 1.5 also retains video-integrated audio generation, with motion, dialogue, and ambient sounds produced in the same process.