Grok Imagine Video 1.5 Preview is xAI’s image-to-video model on VdoBloom, and its defining feature is per-second duration control: any whole-second length from 1 to 15, at 480p, 720p or 1080p, across seven aspect ratios. Pricing scales cleanly with both — roughly 1.5 credits per second at 480p, 3 per second at 720p and 5 per second at 1080p — so a 5-second 720p clip costs 15 credits, or $0.75. This guide covers what it does well, the exact numbers, how to prompt it, and when another model is the better call.

What it is best at

Almost every image-to-video model on the market locks you into two or three preset lengths, usually 5 and 10 seconds. Grok Imagine Video 1.5 Preview does not. You choose 1, 2, 3 and so on up to 15 seconds, which matters enormously the moment you are cutting to music, matching a voiceover line, or filling an exact gap in a timeline. A clip that is genuinely 7 seconds long drops straight into the edit; a 10-second clip you trim to 7 wastes credits on frames you delete.

The second strength is the resolution ladder. Because 480p costs roughly half of 720p and less than a third of 1080p, the model supports a real draft-then-finish workflow: generate eight motion ideas at 480p for about 8 credits each at five seconds, pick the one that works, and rerun only that one at 1080p.

It is image-to-video only. There is no text-to-video mode — the platform routes it exclusively to the image-to-video tab, so you always start from a still. It is rated 5.0 on VdoBloom, the top mark in the model directory, and it is labelled HOT, the flexible content tier, so it tolerates swimwear, dance and fitness subjects that stricter models refuse. Explicit and illegal content remains blocked.

Real specs

Spec	Grok Imagine Video 1.5 Preview
Capability	Image to video only
Durations	Every whole second, 1 to 15
Resolutions	480p, 720p, 1080p
Aspect ratios	1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3
Rating	5.0
Content tier	HOT (flexible)
Provider	xAI

The seven ratios are worth a moment. Along with the obvious 16:9 and 9:16, the list includes 3:2 and 2:3 — the native orientations of most camera sensors. If you are animating real photography rather than an AI-generated still, you can keep the original framing instead of cropping into it.

Exact credit cost per tier

Dollar figures below use the Lite subscription rate, where $15 buys 300 credits, so one credit is $0.05. These credit numbers are read directly from the platform’s pricing table, not estimated.

Duration	480p credits	480p cost	720p credits	720p cost	1080p credits	1080p cost
1 second	2	$0.10	3	$0.15	5	$0.25
3 seconds	5	$0.25	9	$0.45	15	$0.75
5 seconds	8	$0.40	15	$0.75	25	$1.25
8 seconds	12	$0.60	24	$1.20	40	$2.00
10 seconds	15	$0.75	30	$1.50	50	$2.50
15 seconds	23	$1.15	45	$2.25	75	$3.75

The pattern is linear and easy to hold in your head: 1.5 credits per second at 480p, 3 at 720p, 5 at 1080p. The 10 free credits every new account receives, with no card, cover a 5-second 480p test clip with change to spare.

How to prompt it well

An image-to-video prompt is a motion brief, not a scene description. The still already establishes the subject, the wardrobe, the lighting and the set — repeating all of that in the prompt wastes the model’s attention. Describe what moves.

Separate subject motion from camera motion. Write both explicitly: subject turns her head slowly toward the window and smiles; camera pushes in gently on a fixed axis. When you leave the camera unstated, the model invents a move, and invented moves are the most common reason a clip feels wrong.

Keep the motion budget proportional to the duration. A 2-second clip holds one action. A 15-second clip needs a sequence, so give it one: three beats, in order, separated by semicolons. Asking for three actions in two seconds produces a sped-up, jittery result.

State what must not change. Adding a clause such as face, outfit and background stay consistent throughout measurably reduces identity drift over longer durations.

Pick the aspect ratio that matches the source still. The model will render whatever ratio you set, but a mismatch between your photo and the output ratio forces a crop or a fill, and neither improves the frame.

If the source photo shows a real person, you must have that person’s consent before generating or publishing the video. This applies to every photo-animation model on the platform without exception.

A worked example

Suppose you have a product-in-hand photo shot at 3:2 and you want a 7-second social clip.

Open the image-to-video tool and select Grok Imagine Video 1.5 Preview.
Upload the still and set the aspect ratio to 3:2 to match the camera original.
Set duration to 7 and resolution to 480p for the draft pass. That costs 11 credits, or $0.55.
Prompt: hand rotates the bottle slowly to reveal the label; steam drifts upward; camera pushes in two inches on a fixed axis; lighting, background and label text stay unchanged throughout.
Run two or three variations at 480p until the motion reads correctly — about 33 credits, or $1.65, for three attempts.
Rerun the winning prompt at 1080p for 35 credits, or $1.75.

Total for three drafts plus one finished 7-second 1080p clip: 68 credits, or $3.40. Going straight to 1080p for all four runs would have cost 140 credits, so the draft pass more than halves the bill.

When to pick a different model

You need text-to-video. Grok Imagine Video 1.5 Preview cannot start from a prompt alone. Generate a still first with Grok Imagine Pro or another image model, then animate it — or choose a text-to-video model instead.

You need a cheaper draft loop. P-Video renders a 5-second 720p clip for 8 credits against Grok’s 15, and its draft mode drops that to 2 credits. For pure motion exploration it is the cheaper sandbox.

You need longer than 15 seconds in one generation. Fifteen seconds is the hard ceiling here; longer sequences mean stitching clips together or moving to a model with longer native output.

You need the absolute top-end fidelity tier. This is a preview build. It is rated 5.0 and it is excellent, but preview models can change behaviour between updates, so for a locked-down client deliverable a stable flagship model is the safer choice.

Browse the full directory at the Grok Imagine Video 1.5 Preview model page, check credit packs and plans on the pricing page, or start without paying anything on the free AI video generator. Paid plans export watermark-free with commercial rights, as explained on the no-watermark page.

Frequently asked questions

How much does a 10-second Grok Imagine Video 1.5 Preview clip cost?

Fifteen credits at 480p ($0.75), 30 credits at 720p ($1.50), or 50 credits at 1080p ($2.50), using the Lite rate of $0.05 per credit.

Can Grok Imagine Video 1.5 Preview do text-to-video?

No. It is routed exclusively to the image-to-video tab, so every generation starts from a still image you upload or generate first.

What is the longest clip it can make?

Fifteen seconds, and every whole second from 1 to 15 is selectable. That per-second granularity is unusual and is the main reason to choose this model for music-synced edits.

What does Preview mean in the name?

It is the early-access build of the xAI 1.5 generation as offered on VdoBloom, so its behaviour may keep being refined. It already holds the directory’s top 5.0 rating.

Is it restrictive about prompts?

It carries the HOT label, the flexible content tier, so it handles swimwear, dance, fitness and fashion subjects that strict-tier models refuse. Explicit and illegal content stays blocked.

Which aspect ratios can I use?

Seven: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2 and 2:3. The 3:2 and 2:3 options match the native orientation of most camera photos.

How to Use Grok Imagine Video 1.5 Preview: Durations, Credits, Prompts