PixVerse AI Video Generator: V6 Max Price and API (2026)

The PixVerse AI video generator turns a still image and a short text prompt into an animated clip, and in 2026 you can reach it two ways: through the PixVerse consumer app, or through its latest model, PixVerse V6 Max Image To Video, over an API. This guide covers both, with the exact per-second pricing, working code, and an honest comparison with the other video models you are probably weighing up.

What is the PixVerse AI video generator?

PixVerse is a generative video company that ships a mobile app, a web platform and a family of video models. The app wraps those models in a friendly interface: you pick a photo, choose a template or type a prompt, and get back a few seconds of motion. Popular templates include hugging effects, 3D rotations and cinematic zooms, which is why so many of the clips you see on social feeds started there.

Under the hood, the PixVerse AI video generator is an image-to-video model. “Image-to-video” means the model takes a picture as the first frame and predicts how that scene would move, guided by your prompt. Most modern video models, PixVerse included, are built on diffusion: they start from random noise and refine it step by step, across many frames at once, until it matches the input image and description. That is why the first frame stays faithful to your photo while the motion is invented.

The current generation is V6. The top tier, V6 Max, is listed on Unicorn API with a release date of October 24, 2026. It accepts these inputs:

• image_urls: the start image (or images) to animate.
• prompt: a description of the motion, camera and mood.
• quality: 360p, 540p, 720p or 1080p.
• duration: clip length in seconds, which drives the price.
• generate_audio_switch: true to generate the clip with sound, false for silent video.
• seed: a number that fixes the random starting point, so reruns with small prompt tweaks tend to stay closer to the previous result.

PixVerse app, APK or API: which one do you need?

Most people searching for the PixVerse AI video generator land on an app store page or an APK mirror. That is the right path if you want to make a handful of clips on your phone. It is the wrong path if you need fifty product videos a week, want to plug video generation into your own site, or want an agent to produce clips for you.

The app and the APK

The official app is on Google Play, and APK mirror sites host copies of specific versions, including older ones that people look for when an update changes a feature they liked. If you sideload an APK, make sure it is the unmodified original from a reputable mirror, and expect to lose automatic updates.

Why you should skip “mod APK” and “premium unlocked” builds

Searches like “PixVerse mod APK premium unlocked without watermark” are common, and the files are easy to find. Do not install them. A mod APK is the original app decompiled, altered by an unknown third party and repackaged with a new signature. You cannot verify what the added code does, it often asks for permissions the real app never needed, and using it can get your account banned. The same goes for “mod IPA” files for iOS. If the free tier of the app is too limited, the honest options are a paid app plan or paying per clip through an API.

The API route

People also search for a “PixVerse AI API key”. With Unicorn API you use one key for PixVerse V6 Max and every other model in the model catalog, priced in credits, so you can switch models without opening new accounts.

	PixVerse app	PixVerse V6 Max via API
Best for	One-off clips, trending templates	Batches, automation, products and apps
Input	Tap and type on your phone	JSON request with image URL and prompt
Cost model	App plan	Credits per second of video, known before you run
Other models	PixVerse only	Same key for Grok, Kling, Vidu, Wan and more
Automation	Manual	Scripts, backends, AI agents

If you want to try the model before writing code, the video playground runs the same models in the browser.

How much does the PixVerse AI video generator cost?

On Unicorn API, PixVerse V6 Max is billed per second of output. The rate depends on two settings: quality and whether audio is on. The final price is the per-second rate multiplied by the duration, rounded up to whole credits.

Quality	Audio	Credits per second	5-second clip	Plan cost ($0.01/credit)	Top-up cost ($0.015/credit)
360p	Off	4.8	24	$0.24	$0.36
360p	On	7.2	36	$0.36	$0.54
540p	Off	7.2	36	$0.36	$0.54
540p	On	8.8	44	$0.44	$0.66
720p	Off	8.8	44	$0.44	$0.66
720p	On	12	60	$0.60	$0.90
1080p	Off	17.6	88	$0.88	$1.32
1080p	On	22.4	112	$1.12	$1.68

A few things stand out. Jumping from 720p to 1080p doubles the silent rate (8.8 to 17.6 credits per second), so 1080p is worth it for final hero shots, not for drafts. Audio adds between 1.6 and 4.8 credits per second depending on quality. And 360p at 4.8 credits per second is cheap enough to test ten prompt variations for the price of about two 1080p clips.

Credits come from monthly plans at $0.01 each (Starter $25, Creator $50, Studio $100, Pro $200 per month) or from top-ups at $0.015 each that never expire. Full details are on the pricing page. Because the price is fixed by your inputs, you can ask for the exact cost of any request before running it with POST /v1/models/price, which the code below does.

Worked example: a week of product clips

Say you run a small shop and want 20 product videos a week from existing photos. You test each photo at 360p silent for 5 seconds (24 credits), keep the best prompt, then render the final at 720p with audio (60 credits). That is 84 credits per product, or 1,680 credits for 20 products: $16.80 of plan credits, which fits inside the Starter plan. Skip the draft step and render straight to 1080p with audio, and the same batch is 2,240 credits with no chance to fix a bad prompt first.

How to call the PixVerse AI video generator from code

Every media model on Unicorn API follows the same four steps: price the request, submit a job, poll until it finishes, download the output. Here is the full loop for the PixVerse AI video generator.

1. Create a key on the API keys page and export it as UNICORN_API_KEY.
2. Host your start image at a public URL (a CDN, S3 bucket or similar).
3. Send the request body to /v1/models/price to confirm the cost.
4. Send the same body to POST /v1/jobs, then poll GET /v1/jobs/{id} until status is "succeeded" and read outputs[0].url.

Python

import os
import time
import requests

API = "https://api.unicornapi.net"
HEADERS = {"Authorization": f"Bearer {os.environ['UNICORN_API_KEY']}"}

body = {
    "model": "pixverse-v6-max-image-to-video",
    "input": {
        "prompt": "Slow push-in on the sneaker, soft studio light sweeping "
                  "across the leather, shallow depth of field",
        "image_urls": ["https://example.com/sneaker.png"],
        "quality": "720p",
        "duration": 5,  # check allowed durations on the model page
        "generate_audio_switch": False,
        "seed": 42,
    },
}

# 1. Check the exact price first
price = requests.post(f"{API}/v1/models/price", json=body, headers=HEADERS, timeout=30)
price.raise_for_status()
print("Price:", price.json())

# 2. Submit the job
resp = requests.post(f"{API}/v1/jobs", json=body, headers=HEADERS, timeout=30)
resp.raise_for_status()
job_id = resp.json()["id"]
print("Job:", job_id)

# 3. Poll until it finishes
deadline = time.time() + 15 * 60
while True:
    job = requests.get(f"{API}/v1/jobs/{job_id}", headers=HEADERS, timeout=30).json()
    status = job["status"]
    if status == "succeeded":
        break
    if status in ("failed", "canceled", "cancelled"):
        raise RuntimeError(f"Job {job_id} ended with status {status}: {job}")
    if time.time() > deadline:
        raise TimeoutError(f"Job {job_id} still {status} after 15 minutes")
    time.sleep(5)

# 4. Download the video
video_url = job["outputs"][0]["url"]
with open("sneaker.mp4", "wb") as f:
    f.write(requests.get(video_url, timeout=120).content)
print("Saved sneaker.mp4 from", video_url)

curl

curl -s https://api.unicornapi.net/v1/jobs \
  -H "Authorization: Bearer $UNICORN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "pixverse-v6-max-image-to-video",
    "input": {
      "prompt": "Slow push-in on the sneaker, soft studio light sweeping across the leather",
      "image_urls": ["https://example.com/sneaker.png"],
      "quality": "360p",
      "duration": 5,
      "generate_audio_switch": false
    }
  }'

# Then poll with the id from the response:
curl -s https://api.unicornapi.net/v1/jobs/JOB_ID \
  -H "Authorization: Bearer $UNICORN_API_KEY"

The request shapes, status values and error codes are documented in the API docs. If you work in Cursor or Claude Code, the MCP setup lets your coding agent submit and poll these jobs for you.

Choosing settings and writing prompts that work

Image-to-video models already know what your scene looks like, because the image tells them. The prompt’s job is to describe what changes. Prompts that re-describe the picture in detail waste words; prompts that describe motion get better results.

A simple prompt formula

• Subject motion: “the dancer spins once and lands facing the camera”.
• Camera motion: “slow dolly in”, “orbit left”, “static locked-off shot”.
• Light and mood: “golden hour light flickering through leaves”.
• Pace: “slow, gentle”, or “fast, energetic”.

Keep to one main action per clip. Asking a 5-second clip to show a character walk in, sit down, pick up a cup and laugh usually produces rushed, smeared motion. Split it into separate clips and cut them together.

Picking quality and audio

Use 360p while you iterate on prompts, 720p for social posts, and 1080p only for the final version that will be seen large. Turn generate_audio_switch on when ambient sound adds something you would not otherwise have, such as waves or street noise. If you will lay music or a voiceover on top in an editor, leave it off.

Use the seed when refining

Once a draft is close, keep its seed and change only one phrase of the prompt. This makes it much easier to tell whether your edit helped, rather than comparing two unrelated random outcomes.

Start images matter more than prompts

The start frame sets composition, lighting and identity for the entire clip. A sharp, well-lit image with the subject clearly separated from the background animates far better than a cluttered phone snapshot. If your source photo is weak, fix it before you spend credits on video.

How PixVerse compares with other AI video generators

The PixVerse AI video generator is one option among many, and the right choice depends on what you start from and how long the clip needs to be. Here is how PixVerse V6 Max compares with the other video models available on the same key, using a 720p clip as the yardstick.

Model	Input	What stands out	720p price
PixVerse V6 Max	Image + prompt	Four quality tiers, optional audio, seed control	8.8 credits/s silent, 12 with audio
Grok Imagine Video 1.5	Image + prompt	Very low cost per second, strong motion, 1 to 15 s; can be busy at peak times	22.5 credits for 5 s
Grok Imagine Video 1.0	Text or image	Text-to-video, clips from 6 to 30 s	27 credits for 6 s
Wan 2.6 Flash	Image + prompt	Optional audio and multi-shot input, 5, 10 or 15 s	20 credits for 5 s silent, 40 with audio
Vidu Q4 Image To Video	Image + prompt	Resolutions up to 2K and 4K, audio input	15.2 credits/s
Vidu Q4 Reference To Video	Reference images and audio	Builds a clip around reference subjects	15.2 credits/s
Kling 4.0 Keyframe To Video	First, last and middle frames	Control where the shot starts and ends	28.5 credits/s
Kling 4.0 Reference To Video	Images, videos, voices	Reference video input lowers the rate	17 credits/s with video_urls, 28.5 without

Practical reading of that table: for cheap, high-volume animation of stills, Grok Imagine Video 1.5 and Wan 2.6 Flash are hard to beat. PixVerse V6 Max sits in the middle on price and gives you finer control over quality and audio. Kling’s keyframe model is the pick when the clip must end on a specific frame, such as a logo lockup, and Vidu Q4 is the one to try when you need output above 1080p.

What about Sora, Canva, CapCut and the rest?

Many people comparing tools also look at a few names outside this list:

• Sora: OpenAI has discontinued Sora, so anyone who relied on the Sora AI video generator needs a replacement. Image-led workflows move over to PixVerse or Grok without much change.
• Canva and CapCut: the Canva AI video generator and the CapCut AI video generator are features inside video editors. They are convenient if you already edit there, but you work through their interface rather than calling a model directly.
• Synthesia: the Synthesia AI video generator focuses on talking-presenter videos from a script, a different job from animating images.
• InVideo: the InVideo AI video generator assembles edited videos with stock footage and voiceover from a prompt or script.
• Adobe Firefly: the Adobe Firefly AI video generator lives inside Adobe’s apps and suits teams already working there.

These are full products with editors attached. A model API is a building block: you get the raw clip and decide what happens next. If you want to put the PixVerse AI video generator inside your own product, the building block is what you need.

Common mistakes with the PixVerse AI video generator

Most wasted credits come from a short list of avoidable errors.

1. Rendering finals first. Drafting at 360p costs about a quarter of 1080p silent, and an even smaller share of 1080p with audio. Test the motion cheaply, then upscale the winner.
2. Forgetting that duration multiplies everything. PixVerse V6 Max bills per second, so a 10-second clip costs twice a 5-second one. Most social clips loop fine at short lengths.
3. Paying for audio you will replace. If your editor will lay music over the clip, generated audio is thrown away.
4. Using a private or expiring image URL. The model fetches your image from image_urls. Signed URLs that expire in a minute, or files behind a login, cause failed jobs.
5. Describing the image instead of the motion. The model can already see the image; tell it what moves and how the camera behaves.
6. Polling without a timeout. Always stop polling after a sensible limit and handle failed statuses, as in the Python example, so a stuck job does not hang your script.
7. Locking into one model. The PixVerse AI video generator is excellent for some shots and not others. Run your hardest image through two or three models from the catalog before committing a whole campaign.
8. Installing modified apps to dodge limits. A mod APK risks your account and your device. Paying per clip through an API is usually cheaper than people expect, as the worked example above shows.

The fastest way to get a feel for the PixVerse AI video generator is to take one good photo, run it at 360p with three different motion prompts, and compare. Once you have a prompt you like, the same request at 720p or 1080p is one field change away, and the price check tells you exactly what it will cost before you press go.