As of November 10, 2026, xAI’s official API pricing lists image generation from $0.04 per image and video generation from $0.02 per second. The exact amount depends on the model and settings: a 2K medium image costs $0.08, while a 15-second 720p video costs $0.45 with Video 1.5 Lite, $1.05 with Grok Imagine Video, or $2.10 with Video 1.5—before any applicable input-image or input-video charges. These are xAI-direct API prices, not app-subscription or reseller prices.
xAI-direct image prices
For image generation, xAI’s current model page identifies grok-imagine-image-2.0 and lists these output rates:
| Resolution and quality | Output price per image |
|---|---|
| 1K, low | $0.04 |
| 1.5K, low | $0.05 |
| 2K, low | $0.06 |
| 1K, medium | $0.06 |
| 1.5K, medium | $0.07 |
| 2K, medium | $0.08 |
Image edits also incur a listed $0.01 per input image charge, in addition to the generated image’s output price. For example, one edit using one source image and producing one 2K medium image totals $0.01 + $0.08 = $0.09. A text-to-image generation with the same 2K medium output is $0.08, because it has no source-image input. xAI describes image output pricing as per image, not per prompt length.
Example budget: 1,000 text-to-image generations at 2K medium cost 1,000 × $0.08 = $80 in listed output charges. If each is instead an edit with one input image, the estimate is 1,000 × ($0.08 + $0.01) = $90.
xAI-direct video prices
Video output is billed by generated second, with rates varying by model and resolution:
| Model | 480p | 720p | 1080p |
|---|---|---|---|
| grok-imagine-video-1.5-lite | $0.02/sec | $0.03/sec | $0.14/sec |
| grok-imagine-video | $0.05/sec | $0.07/sec | Not listed |
| grok-imagine-video-1.5 | $0.08/sec | $0.14/sec | $0.25/sec |
A 15-second output therefore costs, for video output alone:
- Video 1.5 Lite, 720p: 15 × $0.03 = $0.45
- Grok Imagine Video, 720p: 15 × $0.07 = $1.05
- Video 1.5, 720p: 15 × $0.14 = $2.10
- Video 1.5, 1080p: 15 × $0.25 = $3.75
The model pages also list input charges where image or video inputs are used. For Grok Imagine Video, listed inputs include $0.002 per image and $0.01 per second of input video. Video 1.5 and 1.5 Lite list image inputs at $0.01 per image. Add applicable input charges to the output arithmetic; don’t assume these output-only examples cover an image-to-video or video-editing request.
Example production estimate: 100 text-to-video clips, each 10 seconds at 720p using Video 1.5 Lite, cost 100 × 10 × $0.03 = $30 in listed video output charges. This assumes the rate shown in xAI’s pricing table applies and excludes charges not represented by that output calculation.
Which model should you budget for?
- Video 1.5 Lite has the lowest listed per-second rates for 480p and 720p. It supports text-to-video and image-to-video with native audio; its 1080p option is upscaled from 720p, not native 1080p. xAI’s video overview recommends it for simple videos. (
docs.x.ai) - Video 1.5 is the option to consider when you need native 1080p or reference-based controls. xAI lists text-to-video and image-to-video at native resolutions up to 1080p, up to 14 reference images, and pinned first, last, and mid-video frames. It also lists up to three voice references; voice references from your own audio files are available to trusted partners on request. (
docs.x.ai) - Grok Imagine Video is the model xAI’s overview lists for video editing and extension. The overview specifies source clips up to 8.7 seconds for editing and extension of 2–10 seconds. Its output duration is listed as 1–15 seconds, as are the other two models; check the operation-specific limits rather than assuming a generation limit also describes editing or extension. (
docs.x.ai)
The often-quoted $4.20 per minute is not a universal Imagine video rate: it equals the listed $0.07/sec rate for Grok Imagine Video at 720p. Other models and resolutions have different rates. xAI’s overview describes native audio for these models, but the pricing table lists video rates rather than a separate audio line; confirm the billing treatment for your chosen route before relying on an audio-inclusive budget.
Model and API details that affect the estimate
For direct image generation, xAI documents POST /v1/images/generations and supports requesting 1–10 images with the n parameter. The current image model page names grok-imagine-image-2.0. Image edits use the separate POST /v1/images/edits endpoint; xAI says edits are billed for both the input image and the generated output. (docs.x.ai)
For video generation, xAI lists output durations of 1–15 seconds for Video 1.5 Lite, Video 1.5, and Grok Imagine Video. Requests are asynchronous: submit a request, poll using the returned request ID, then download the clip. The xAI SDK can handle polling automatically. The 15-second figure is a generation-duration limit in the overview; Grok Imagine Video’s editing and extension operations have their own listed source-clip and extension ranges. (docs.x.ai)
One version detail matters: xAI’s migration notice says that from November 2, 2026, requests using the grok-imagine-image-quality slug are served by grok-imagine-image-2.0 at low quality. The notice advises specifying the 2.0 model and quality explicitly to control which tier is used and billed.
Quick budget formulas
- Image generation: number of generated images × the chosen resolution/quality output rate.
- Image editing: generated-image output cost + $0.01 × the number of input images, using the listed Image 2.0 input rate.
- Video generation: generated seconds × the selected model-and-resolution rate, plus applicable image or video input charges.
These calculations use xAI’s published USD rates. The cited pricing pages do not settle every production billing question, such as how failed requests, client timeouts, or retries are charged. For a launch budget, test the exact model and request type you plan to deploy, then reconcile the returned usage cost with the estimate before increasing volume.
Sources
Grok Imagine Image 2.0 | SpaceXSI Docs
Grok Imagine Video | SpaceXSI Docs
Grok Imagine Video 1.5 | SpaceXSI Docs
Grok Imagine Video 1.5 Lite | SpaceXSI Docs
Video Overview: Grok Imagine Video API | SpaceXSI Docs
grok-imagine-image-quality Retirement on November 2, 2026
Image Generation | SpaceXSI Docs