Grok Imagine V1.5 Lands on ZOOOP: xAI's Image-to-Video Finally Holds Steady
Grok Imagine V1.5 is now live on ZOOOP. It's xAI's latest take on image-to-video, and compared to 1.0 the real wins are simple: the picture holds together, and the motion looks more natural. In preview it even topped the public image-to-video leaderboard. Here's what's new and how to start using it.
What Grok Imagine V1.5 actually does
First, the scope. Grok Imagine V1.5 is a dedicated image-to-video model — you hand it a still image as the starting point, write a line describing the motion, and it animates that frame into a short clip. It doesn't generate stills and it doesn't do text-to-video; it does one thing, "make this image move." You'll find it inside AI video generation on ZOOOP.
It comes from the same line as the original Grok Imagine, so the foundation carries over. What this version really fixes are the rough edges that bugged people most last time.
The big one: the picture finally holds steady
The original Grok Imagine got the most flak for stability — faces drifted, subjects wandered, scene elements changed frame to frame. Steadiness is exactly what V1.5 improves: across the whole clip, the subject, the face, and the scene stay consistent instead of warping or sliding around.
For image-to-video this matters more than raw prettiness. A clip that morphs is unusable no matter how sharp each frame looks; one that stays coherent from start to finish is one you can actually ship. The jump Grok Imagine V1.5 makes here is easy to see with your own eyes.
Better motion, and a leaderboard win
Beyond stability, the motion quality is better too — how faithfully a single still carries into a moving shot is handled more convincingly than in 1.0, with less of that "fake" look.
To be fair about it: Grok Imagine V1.5 reached first place on the public Image-to-Video Arena during its preview. A leaderboard is just one signal, and it doesn't mean the model wins everywhere — but it does say this version sits in the top tier at turning a still into a moving shot.
Sound still comes baked in
Worth a note: audio isn't new here. Since the previous generation, Grok Imagine has shipped video with sound generated alongside the picture — dialogue, ambient noise, sound effects, with lip-sync on talking characters. The latest version simply keeps it.
The payoff is still real. Making a short clip with sound used to mean three passes — generate the picture, add a voice with a separate tool, then handle effects on top. Here a single prompt gets you a clip that already has sound, which saves time and the hassle of lining up outputs from different tools.
Specs at a glance: what it can and can't do
A few practical numbers:
- Output at 720p or 480p
- 1 to 15 seconds, 5 seconds by default
- Image-to-video only — you need a starting image first
The sweet spot is clear: animate a landscape or a product shot with ambient sound, make vertical clips for social, or bring a concept frame to life just to check the pacing.
The limits are worth stating too. Grok Imagine V1.5 caps at 720p — it's not a 1080p / 4K finishing model; it animates a single shot, not multi-shot cuts; and it won't conjure an image out of nothing. If you don't have a frame to animate yet, generate one with AI image generation first and feed it in. For higher resolution or multi-shot edits, heavier video models like Kling V3, Seedance V2.0, Veo 3.1… are the better target.
What it means for creators
Do the math and it clicks. Making a short voiced clip used to mean bouncing between generating footage, dubbing, and adding effects — and worse, one warped frame could sink the whole take. Grok Imagine V1.5 shortens that chain and shores up the stability, which is real time saved for anyone pushing out a lot of short clips with sound.
It's not a do-everything model. Its sweet spot is "make this image move — steady, and with sound": talking heads, product motion, social clips, quick previews. When you need a high-res final or complex editing, hand the shot to another model. That division of labor is part of why Grok Imagine V1.5 is so usable in the lane it's built for.
How to start on ZOOOP
In AI video generation on ZOOOP, pick Grok Imagine V1.5, upload a starting image (it becomes the first frame), write a prompt describing the motion, set resolution and duration, and generate.
Every model on ZOOOP shares the same credits, and credits never expire — so you can bounce between models without worrying about a wasted pick. To dig into the parameters and examples first, see the Grok Imagine V1.5 model page.