xAI drops Grok Imagine 1.5 Preview for developers

Just as the video AI space saw a major heavyweight step aside, xAI is jumping right into the ring.

Elon Musk's xAI has rolled out Grok Imagine 1.5 Preview, also referred to as Grok Imagine Video 1.5. The new image-to-video model is now officially available to developers as an early preview through the xAI API.

Still photos turned into 720p motion

Basically, the model takes a single still image and animates it into a short video at up to 720p resolution.

Users can steer the output using text prompts that describe specific camera movements, pacing, and overall atmosphere. According to xAI, the engine animates the scene while keeping the original image's fine details and lighting conditions intact.

Another key capability is shot stitching. Developers can string multiple clips together into longer sequences while maintaining visual consistency across the entire scene.

Straight to the API

Instead of launching a flashy consumer web app first, xAI went straight to developers. The official Grok account announced the release on X with a direct link to the API endpoint.

According to reports, developers can hook the model into their own pipelines with just a few lines of code. The launch puts xAI in direct competition with dedicated video AI platforms like Seedance and Google's Veo.

The timing is especially striking. OpenAI recently pulled its high-profile Sora video model from circulation, citing computing resource constraints and what industry watchers widely viewed as an unsustainable business model.

Why it matters

We think xAI's decision to drop Grok Imagine 1.5 straight into an API preview is a clever, opportunistic grab for developers left stranded by Sora's sudden retreat.

For builders, animators, and independent creators, having an image-to-video tool that responds to camera prompts and supports multi-shot consistency is immediately useful for storyboarding, dynamic ads, and game asset prototyping. It can be set up in minutes with minimal code. That said, capping the resolution at 720p means it is strictly a prototyping and short-form tool for now, not a replacement for high-end production pipelines.

For the wider industry, this release marks a reality check in the AI video race. Full text-to-video generation has proven brutally expensive to run and difficult to monetize, as OpenAI's pullback showed. By narrowing the scope to image-to-video, focusing on controlled camera work, and billing directly through an API, xAI is testing whether a more constrained model can actually sustain itself.

The smart part of the strategy is using image conditioning to keep consistency high while keeping generation targets at 720p to control compute costs. The lingering question, however, is whether xAI can stand its ground against dedicated competitors like Seedance and deep-pocketed giants like Google's Veo without running into the exact same compute bottlenecks that forced OpenAI to tap out.