Grok Imagine Pricing by Resolution and Duration
> A plain reading of the Grok Imagine price sheet: 480p at $0.05/s, 720p at $0.07/s, edit-video at $0.06 to $0.08/s, images at $0.02 per output. With per-minute math for social runs.
The Grok Imagine pricing surface is small once you list it out, but the combinations matter. Resolution and duration drive video cost, edit-video has its own scale, and images split between generation and edit. Nothing here is tiered by tokens or prompt length. You pay for seconds and pixels.

The base table
| Endpoint | Resolution | Per-second rate |
|---|---|---|
| `xai/grok-imagine-video/text-to-video` | 480p | $0.05 |
| `xai/grok-imagine-video/text-to-video` | 720p | $0.07 |
| `xai/grok-imagine-video/image-to-video` | 480p | $0.05 + $0.002 per input image |
| `xai/grok-imagine-video/image-to-video` | 720p | $0.07 + $0.002 per input image |
| `xai/grok-imagine-video/edit-video` | 480p (capped) | $0.06 to $0.08 per second |
| `xai/grok-imagine-image` | 1k or 2k | $0.02 per image |
| `xai/grok-imagine-image/edit` | 1k or 2k | $0.02 + $0.002 = $0.022 |
Audio is included. No audio surcharge. Duration ranges from 1 to 15 seconds per clip. Frame rate is locked at 24 fps. You cannot buy a 16-second clip or drop to 12 fps to cut cost. Need 30 seconds? Run two 15-second calls and stitch.
Example math
10 seconds at 720p text-to-video: 10 x $0.07 = $0.70 per clip. The most common unit in a Grok run. Memorize this one.
10 seconds at 480p: 10 x $0.05 = $0.50. For internal previews and storyboard drafts this is almost always the right call.
15 seconds at 720p: 15 x $0.07 = $1.05. The longest clip Grok will produce in one call. If you do not need every frame, cut to 10s and save $0.35.
8 seconds at 480p image-to-video: 8 x $0.05 + $0.002 = $0.402. Trivial per clip, non-trivial across a batch.
6 seconds edit-video at the $0.07 midpoint: $0.42. Exact rate ranges $0.06 to $0.08. Budget $0.07 unless you know otherwise.

Per-minute math for social runs
Most teams buy minutes, not clips. Here is the cost to produce 60 seconds of finished video, assuming you land every clip on the first try (add 20% for retries).
| Output | Res | Clip | Count | Cost |
|---|---|---|---|---|
| 60s hero reel | 720p | 10s | 6 | $4.20 |
| 60s hero reel | 480p | 10s | 6 | $3.00 |
| 60s TikTok vertical | 720p | 8s | 8 | $4.48 |
| 60s TikTok vertical | 480p | 8s | 8 | $3.20 |
| 60s retargeting | 480p | 5s | 12 | $3.00 |
A weekly batch of ten 60s 720p hero reels is $42 before retries, $50.40 with the buffer. $2,620/year.
Edit-video pricing
Edit-video sits between $0.06 and $0.08 per second. The endpoint caps at 854x480 and truncates to 8 seconds, so there is no resolution switch to price against. Instead, rate varies by operation. Recolor and style shifts hit the floor. Full object replacement or temporal mask edits hit the ceiling.
- 8s edit at $0.06: $0.48.
- 8s edit at $0.08: $0.64.
- Midpoint: $0.56.
Truncation is silent. Pass a 15s source and you get 8s back.
Images
One call produces up to four images at 1k or 2k. You pay $0.02 per image, so a four-image call is $0.08. Edit adds $0.002 input on top of the $0.02 output: $0.022 per edited image.
1import { fal } from "@fal-ai/client";23fal.config({ credentials: process.env.FAL_KEY });45const batch = await fal.subscribe("xai/grok-imagine-image", {6 input: {7 prompt: "editorial shot of a ceramic studio workbench, warm afternoon light, detailed texture on wheel and clay",8 num_images: 4,9 image_size: "2k"10 },11 logs: true12});1314for (const image of batch.data.images) {15 console.log(image.url);16}
Four 2k images for $0.08 is aggressive. If you plan to pick a favorite from four, the marginal cost of the other three is $0.06, often worth it to dodge a regeneration.
A realistic weekly run
- 10 x 10s 720p hero clips at $0.70 = $7.00
- 30 x 8s 480p TikTok cutdowns at $0.40 = $12.00
- 40 x 2k images at $0.02 = $0.80
- 5 x 8s edit-video passes at $0.56 = $2.80
Total: $22.60/week. $1,175/year. No minimums, no subscription, no tier gates. You pay for what you render.