xAI Grok image-to-video generative video AI media

Grok Imagine 1.5 Arrives With Image-to-Video at 720p

5 June 2026

xAI has released Grok Imagine Video 1.5 in preview, an image-to-video model that animates still images into short clips at up to 720p resolution using text prompts to control camera movement, pacing, and atmosphere. The model is available via the xAI API and also supports stitching multiple shots into longer, visually consistent scenes. This puts xAI in direct competition with video AI providers such as Seedance and Google's Veo.

xAI has quietly pushed out Grok Imagine Video 1.5 in preview, adding image-to-video generation to its growing roster of media tools. The model takes a single static image and turns it into a short clip at up to 720p resolution, with users steering the output through text prompts covering camera movement, pacing, and mood.

The results apparently respect the source image's original lighting and composition rather than drifting into generic AI-video territory. Multiple shots can be chained together for longer sequences, with xAI claiming consistent visual style across cuts.

For now it's API-only, which means developers and tinkerers rather than casual users. A few lines of code is all it supposedly takes to get running.

The timing is interesting. The image-to-video space is getting crowded fast. Seedance and Google's Veo are already established there, and OpenAI quietly shelved Sora not long ago, blaming resource constraints. The real issue was probably that nobody had worked out how to actually make money from it.

xAI clearly thinks it has a shot at carving out space in a market that, despite the hype, still hasn't produced a clear winner.