AI Models
What each model is good at, where it differs from the others, and what it costs to run.
xAI's unversioned Grok stills model. Fast 9:16 tries — not Grok Imagine 2.0, and not a video row.
Credits come with a Generator plan. The exact cost is shown on the Generate button before anything is spent.
Grok Imagine is xAI's unversioned stills SKU — the cheap Grok picture, not Grok Imagine 2.0, and not a clip. On Visxy it is a throwaway pass through the image prompt library: 9:16, one idea, speed over a file you will inspect.
If you already know the still is a keeper, you are on 2.0. If you wanted motion, that is Grok Imagine Video or Grok Imagine 1.5.
Trying looks. Paste a Visxy 9:16 portrait and see whether the pose and the crop read. Effort on this row is a speed-versus-quality pass, not a different Grok. Draft on the faster mark; only spend the heavier one when the crop is already the one you want.
What it is not: 2.0 with a smaller slider, grok.com's Fun/Spicy playground, or a 1K/2K print engine. Those chips are not how this Visxy row is sold.
Against Grok Imagine 2.0: different catalog row, different family id. This is the unversioned stills SKU. 2.0 is the current stills model. You do not "turn quality up" into 2.0.
Against the video rows: this page does not move. Animate a still on the video pages, or start from a prompt there.
This row can edit. 2.0 cannot, on Visxy — kie's 2.0 editor wants a mask and a prior task id we do not keep.
Paste the library prompt. Stay on 9:16. One idea.
If the result is only close, re-run the same text on 2.0 rather than rewriting it for this page.
No Visxy still is published yet. Generate fewer.
There is no public Grok Imagine Edit page. Attach one still. There is no aspect picker on that shape — crop first. Describe the change, not a new brief.
No. This is the unversioned stills SKU. 2.0 is its own page.
No. Use Grok Imagine Video or Grok Imagine 1.5.
No. Attach one picture to Grok Imagine and describe the change.
