The Grok Imagine Price in 2026: Free, Plans and API Costs

The Grok Imagine price you pay depends on where you use it: inside xAI’s Grok app on a subscription, or through an API that bills you per image and per second of video. Those are two very different deals. The app bundles Imagine with chat and voice under tiered plans that xAI changes often, while the API route charges a fixed, visible amount for each generation and nothing when you are idle.

This guide explains both, gives exact per-request costs for the API route, works through a realistic monthly budget, and shows the code to generate images and videos. If you only need the short version: casual users who already pay for Grok should use the app, and anyone building a product, a batch workflow or a content pipeline will find per-request billing cheaper and far more predictable.

What does Grok Imagine cost in the Grok app?

Grok Imagine is xAI’s image and video generator. In the consumer product it lives inside the Grok app and on X, where you type a prompt, get a grid of images, and can animate one of them into a short clip. You do not buy Imagine on its own there: it is one feature of the Grok subscription, alongside chat, voice and coding tools.

That bundling is why the Grok Imagine price is hard to pin down from the app side. Three things make it slippery:

• Tiers change. xAI has added, renamed and repriced Grok plans several times, and which tier unlocks which Imagine features (longer videos, higher resolution, faster queues) has moved with them. Check the plan screen inside the app for the current numbers rather than trusting a screenshot from a few months ago.
• Usage limits are soft. Subscriptions come with generation limits and rate limits, and xAI does not always publish them precisely. Heavy users report hitting caps during busy periods even on paid plans.
• Free access is not stable. The steady search volume for phrases like “grok imagine no longer free” tells the story: free Imagine usage has been expanded, restricted and reshaped repeatedly. If you are asking “is Grok Imagine free again yet?”, the answer genuinely depends on the week.

Who the app subscription suits

If you already pay for Grok because you use the chatbot daily, Imagine is effectively included and the marginal cost of a few images is zero. For that person, the app is the right choice. The calculation flips when Imagine is the main thing you want. Paying a monthly subscription to make a handful of videos is expensive per clip, and paying it to make thousands runs into limits you cannot see in advance.

The app also has no way to plug into your own software. You cannot call it from a script, batch 300 product shots overnight, or wire it into a Next.js backend. For any of that you need API access, which is where per-request pricing comes in.

What is the Grok Imagine price per image and video on the API?

On Unicorn API, every model is priced in credits, and the Grok Imagine price is set per request with no subscription to xAI. Credits cost $0.01 each on monthly plans (Starter $25, Creator $50, Studio $100, Pro $200 per month) and $0.015 each as top-ups that never expire. You can also ask the API for the exact price of any request before you run it, using POST /v1/models/price.

The Grok family on the platform is four models:

Model	What it does	Price in credits
Grok Imagine	Text-to-image; returns 6 variations per text prompt (2 when reference images are attached)	4 per request by default; multiplied by n when set
Grok Imagine Video 1.5	Image-to-video only, 1 to 15 seconds	2.4/s at 480p, 4.5/s at 720p, 8/s at 1080p
Grok Imagine Video 1.0	Text-to-video or image-to-video, 6 to 30 seconds	Same per-second rates as 1.5
Grok Imagine Upscale	Upscales an earlier generation, referenced by task_id	10 at 720p, 20 at 1080p

Images

A default Grok Imagine image request costs 4 credits: about $0.04 on a monthly plan or $0.06 from top-up credits. The model is built to give you a spread of options in one go, six variations from a text prompt or two when you attach reference images. The n field multiplies the price, so only set it when you deliberately want more than one batch. Supported aspect ratios are 1:1, 2:3, 3:2, 16:9 and 9:16.

Video

Video is billed per second of output, and resolution is the main lever. Here is what common clip lengths cost with Grok Imagine Video 1.5:

Clip	480p	720p	1080p
5 seconds	12 credits ($0.12)	23 credits ($0.23)	40 credits ($0.40)
10 seconds	24 credits ($0.24)	45 credits ($0.45)	80 credits ($0.80)
15 seconds	36 credits ($0.36)	68 credits ($0.68)	120 credits ($1.20)

Dollar figures use the $0.01 monthly-plan rate; multiply by 1.5 for top-up credits. Note the rounding: each output is rounded up to whole credits, so a 5-second 720p clip listed at 22.5 credits is billed as 23. At 720p, even-numbered durations land on whole numbers and waste nothing.

Video 1.0 uses the same per-second rate but covers 6 to 30 seconds and accepts a plain text prompt with no start frame. A 30-second clip costs 72 credits at 480p, 135 at 720p and 240 at 1080p. Use 1.0 when you need length or text-to-video, and 1.5 when you have a start image and want its stronger motion quality. Version 1.5 is a preview and can be busy at peak times, so build retries into anything time-sensitive.

Is spicy mode part of the Grok Imagine price?

“Grok Imagine spicy mode” is one of the most searched phrases around this product, so it deserves a straight answer. Spicy mode is a content setting in xAI’s own app. It is governed by xAI’s account rules, age checks and regional restrictions, and it is not a separately priced add-on you buy with a card. Third-party resellers that advertise it bundled into a credit pack are describing access through their own accounts, which is worth treating with caution.

On the API side there is no app-style toggle. Requests run under the model provider’s content policy plus the platform’s terms of use, and those apply equally to every customer. The practical advice for builders:

• Do not design a product around app-only settings carrying over to an API. They usually do not, and policies can tighten without notice.
• Expect some prompts to be refused, and handle a failed job gracefully in your code rather than retrying the same prompt in a loop.
• Check the model page and the documentation for the current content rules before you commit to a use case.

So spicy mode does not change the Grok Imagine price at all. It changes what the app will generate, not what anything costs.

Worked example: a month of Grok Imagine on a budget

Abstract per-second rates are hard to reason about, so here is a realistic month for a small team making social content: product images, short animated clips, and a few polished hero videos.

1. Images: 200 Grok Imagine requests to explore concepts and pick start frames. 200 × 4 = 800 credits.
2. Video: 60 clips of 8 seconds at 720p with Video 1.5, animating the best frames. Each costs 36 credits, so 60 × 36 = 2,160 credits.
3. Upscales: the 20 best clips upscaled to 1080p with Grok Imagine Upscale. 20 × 20 = 400 credits.

Total: 3,360 credits. The Creator plan at $50 a month gives 5,000 credits, so this workload fits with 1,640 credits to spare for re-rolls. The same work bought with top-up credits would cost about $50.40, so the monthly plan only wins if you use most of its credits; light or irregular users can stick with top-ups, which never expire.

Generate at 1080p or upscale later?

This example hides a useful trade-off. An 8-second clip generated directly at 1080p costs 64 credits. Generating it at 720p (36 credits) and upscaling only if it turns out well (20 credits) costs 56 credits for keepers, and just 36 for the clips you reject. When most of your drafts get thrown away, which is normal with video, the generate-small-then-upscale pattern saves a meaningful share of the bill. Whether an upscaled clip matches native 1080p quality for your footage is something to test on a few samples before you standardise on it.

The bigger lesson: with per-request billing, the Grok Imagine price you actually pay is driven by your rejection rate. Cheap drafts at 480p, a tight shortlist, and high resolution only for the finalists will stretch any budget further than switching plans.

How to use the Grok Imagine API with code

All media models on Unicorn API follow the same pattern: create a job with POST /v1/jobs, poll GET /v1/jobs/{id} until its status is "succeeded", then read the file URL from job.outputs[0].url. First, create an API key and export it as UNICORN_API_KEY.

Before running anything, you can check a request’s cost. This asks what a 10-second 720p clip will be billed at:

curl -X POST https://api.unicornapi.net/v1/models/price \
  -H "Authorization: Bearer $UNICORN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"grok-imagine-video-1-5-preview","input":{"prompt":"Slow push-in, steam rising","duration":10,"resolution":"720p"}}'

The Python script below does the full pipeline: generate images with Grok Imagine, take the first output as a start frame, animate it with Video 1.5, and print the price of each step before it runs.

import os
import time
import requests

BASE = "https://api.unicornapi.net"
HEADERS = {
    "Authorization": f"Bearer {os.environ['UNICORN_API_KEY']}",
    "Content-Type": "application/json",
}


def get_price(body):
    r = requests.post(f"{BASE}/v1/models/price", json=body, headers=HEADERS)
    r.raise_for_status()
    return r.json()


def run_job(body, poll_seconds=3):
    r = requests.post(f"{BASE}/v1/jobs", json=body, headers=HEADERS)
    r.raise_for_status()
    job = r.json()
    while job["status"] not in ("succeeded", "failed"):
        time.sleep(poll_seconds)
        r = requests.get(f"{BASE}/v1/jobs/{job['id']}", headers=HEADERS)
        r.raise_for_status()
        job = r.json()
    if job["status"] != "succeeded":
        raise RuntimeError(f"Job {job['id']} did not succeed: {job}")
    return job


# Step 1: images with Grok Imagine (4 credits by default)
image_body = {
    "model": "grok-imagine",
    "input": {
        "prompt": "A ceramic coffee mug on a wooden table, morning window light, steam rising, product photo",
        "aspect_ratio": "16:9",
    },
}
print("Image price:", get_price(image_body))
image_job = run_job(image_body)
start_frame = image_job["outputs"][0]["url"]
print("Start frame:", start_frame)

# Step 2: animate it with Grok Imagine Video 1.5 (10 s at 720p = 45 credits)
video_body = {
    "model": "grok-imagine-video-1-5-preview",
    "input": {
        "image": start_frame,
        "prompt": "Slow camera push-in, steam curling upward, soft light shifting",
        "duration": 10,
        "resolution": "720p",
        "aspect_ratio": "16:9",
    },
}
print("Video price:", get_price(video_body))
video_job = run_job(video_body)
print("Video:", video_job["outputs"][0]["url"])

A few notes on the code. Video 1.5 is image-to-video only, so the image field is required in practice; if you want to start from text alone, swap the model id to grok-imagine/image-to-video and use a duration of 6 to 30 seconds. Because 1.5 can be busy at peak times, production code should catch the error from run_job and retry later rather than immediately. If you would rather test prompts by hand first, the browser playground runs the same models without writing any code, and you can call the API from Cursor or Claude Code through the MCP integration.

Grok Imagine alternatives: how the price compares

A common reason people search for a Grok Imagine alternative is that the app’s limits or free tier changed under them. On the API side, switching models is a one-line change to the model field, so it is worth knowing what sits next to Grok at similar prices.

Model	Price per request	Strengths	Reference images
Grok Imagine	4 credits	Several variations per prompt, fast concept exploration	Yes (returns 2 variations)
Nano Banana 2 Lite	4 credits	Cheapest Gemini image model, fast 1K images and edits, extra aspect ratios including 21:9 and 4:3	Up to 10
GPT Image 1.5	3 low, 4 medium, 22 high	Prompt adherence, legible text, optional transparent background	Up to 4

How to choose:

• Exploring ideas: Grok Imagine. Getting a spread of variations per request is ideal for picking a direction or a start frame for video.
• Editing and consistency: Nano Banana 2 Lite. Ten reference images make it the better pick for keeping a character or product consistent across shots.
• Text in the image: GPT Image 1.5. Labels, posters and packaging mockups with readable text are its strength, and the transparent background option helps with product cut-outs.

For a deeper look at the higher-end image models, see GPT Image 2 vs Nano Banana Pro. If you are weighing Grok’s video models against other providers, our Runway text-to-video guide covers a different set of trade-offs, and the full model catalog lists every option with live prices.

Mistakes that make the Grok Imagine price higher than it should be

Most overspending on generative media comes from a few habits rather than from the rate card. These are the ones that show up most often.

1. Rendering drafts at 1080p. 1080p costs about 3x as much per second as 480p. Use 480p to check motion and composition, then re-render or upscale only the clips you keep.
2. Setting n without meaning to. On Grok Imagine, n multiplies the request price. Copying an example that sets n to 4 quietly makes every call 16 credits.
3. Odd durations at 720p. Rounding up to whole credits means 5, 7 or 9 seconds at 720p each waste half a credit. It is small per clip and adds up over thousands.
4. Using Video 1.0 for short clips. Video 1.0 starts at 6 seconds. If you only need 3 seconds and have a start frame, Video 1.5 costs 13.5 credits at 720p (billed as 14) against 27 for the shortest 1.0 clip.
5. Retrying refused prompts in a loop. If a job fails on content grounds, the same prompt will usually fail again. Log it and change the prompt.
6. Skipping the price check. Calling /v1/models/price before a batch job is cheap insurance, especially when durations and resolutions come from user input.
7. Buying the wrong credit type. Monthly plans are cheaper per credit but reset each month; top-ups cost more per credit but never expire. Match the option to how steady your usage is.

Get those right and the Grok Imagine price becomes something you can forecast in a spreadsheet: a fixed number of credits per image, a fixed rate per second of video, and a visible quote before every request. That predictability is the real difference from app subscriptions, where the limits move and the per-generation cost is something you can only estimate. The full rate card for every model is on the pricing page.