What is new in Kling 4.0 compared with Kling 3.0?

Longer clips (up to 30 seconds instead of 15), a 21:9 ultrawide option, in-between keyframes on image to video, and reference to video that takes videos and voice clips as well as images. Sound is always on. Kling 3.0 still offers 4K and Motion Control, which Kling 4.0 does not have here.

Can I turn the audio off?

No. Every Kling 4.0 clip comes back with a stereo soundtrack, at no extra charge. On reference to video you can keep a reference video's own sound instead with keep_original_audio.

How do I refer to my uploads in the prompt?

On reference to video, write @Image1, @Image2 for images, @Video1 for videos and @Audio1 for audio clips, each numbered in the order you uploaded them. In-between keyframes on image to video are not named in the prompt; their order in the list is the order they appear.

Does Kling 4.0 allow NSFW content?

No. Requests are screened before generation and everything from the Moderate level up is refused within a few seconds; a refused request is not charged. For that kind of work, browse the unrestricted video families in our catalogue instead.

Per second of output. 1080p costs more per second than 720p. On reference to video the seconds of any reference video are added at a lower per-second rate. Aspect ratio, keyframes, reference images and audio clips do not change the price.

How long does a clip take?

In our tests a 3-second clip took about 1.5 to 4 minutes and a 5-second clip a few minutes more. Requests on the same endpoint are processed one at a time.

Kling 4.0 API: 30 s Video, Keyframes, References

kling/4.0/text-to-video

Kling 4.0 text-to-video turns a written brief into a clip of 3 to 30 seconds at 720p or 1080p. Choose 16:9, 9:16, 1:1 or 21:9 ultrawide; every clip comes with a stereo soundtrack.

From $0.216/s

Run model View API

Input

Text to video

Prompt

Aspect ratio

Resolutionaffects price

Durationaffects price

Safety checkerOff

SpicyAPI does not run a safety checker on this model. Your prompt and media go to the model as written, and the model itself decides what comes back.

Checking your account before enabling submission and pricing.

Output

Ready for a new run

Example result

Extend

Video upscaling

Video editing

{
  "model": "kling/4.0/text-to-video",
  "input": {
    "aspect_ratio": "16:9",
    "resolution": "720p",
    "duration_seconds": 5,
    "prompt": "Inside a white observatory dome on a high-desert mesa at indigo dusk. An astronomer in her late thirties with short silver-streaked hair, wearing a charcoal wool sweater and fingerless gloves, turns a heavy brass wheel; with a deep rumble the dome slit slides open and a band of deep-blue sky full of first stars pours light across the polished chrome telescope. She looks up, breath visible in the cold air, and says quietly, \"There you are.\" Slow crane up from her hands on the wheel to the opening sky. Sound: the grinding rumble of the dome motor, desert wind, the soft hum of equipment."
  }
}

Introduction

Kling 4.0 Text to video turns a written brief into a clip of 3 to 30 seconds at 720p or 1080p, in 16:9, 9:16, 1:1 or 21:9 ultrawide. Every clip comes back with a stereo soundtrack generated in the same pass: dialogue you write in quotes, effects and ambience. There is no audio switch and no extra charge for it.

Choose a Kling 4.0 mode by starting material

What you have at the start	Mode
a written scene or script	Kling 4.0 Text to video (you are here)
an opening frame, and optionally a closing frame and in-between frames	Kling 4.0 Image to video
photos, clips or voice samples to cast from	Kling 4.0 Reference to video

Key features & highlights

Up to 30 seconds in one take

Lengths run from 3 to 30 whole seconds, twice the ceiling of Kling 3.0. A long take lets a scene breathe - a slow crane, a full walk across a set - without stitching shorter clips together.

Sound is always on

Each clip carries a two-channel soundtrack. In our tests the dome motor's rumble, wind and a whispered line all came back from the prompt's Sound: line, timed to the picture.

21:9 ultrawide

Besides 16:9, 9:16 and 1:1, Kling 4.0 renders a true 21:9 frame. Our 1080p ultrawide test came back at 2200x940.

Request parameters

Request parameters for Kling 4.0 · Text to Video
Parameter	Required	Description
`prompt`	Yes	Describe the subject, the action, the camera and the sound. Put spoken lines in quotes and the sound on its own line. at least 1 characters
`aspect_ratio`	No	16:9 (default), 9:16, 1:1 or 21:9 ultrawide. It does not change the price. `16:9` · `9:16` · `1:1` · `21:9` · Default: `16:9`
`resolution`	No	720p or 1080p (default 720p). 1080p costs more per second. `720p` · `1080p` · Default: `720p`
`duration_seconds`	No	Whole seconds, 3 to 30 (default 5). Drives the price. `≥ 3` · `≤ 30` · Default: `5`

Pricing

Billed per second of output. 1080p costs more per second than 720p. Every clip comes with a stereo soundtrack at no extra charge. A refused or failed request is not charged.

Pricing
Specification	Price
`720p`	$0.216 / second
`1080p`	$0.324 / second

Use cases

Establishing shots with sound

A landscape, a building, a road at dusk - write the place, the light, one camera move and the ambience, and the clip is usable as an opener with its sound already in place.

Short scenes with a spoken line

Put the line in quotes after the action. One clear line per clip is easier for the model to time than overlapping dialogue.

Vertical clips for social

Choose 9:16 and frame the subject in the centre third; the model composes for the vertical frame instead of cropping a wide shot.

Explore key use-case tutorials

Prompt guide & pro tips

Write it in the order it happens

Subject and wardrobe, action, camera move, look, then a separate Sound: line. Our observatory example follows exactly that order.

Draft short, then go long

Generation is slow - a 3-second test took two to four minutes, and requests run one after another rather than in parallel. Find the framing with 3 to 5 seconds at 720p, then render the long or 1080p version.

Describe the sound you want

Because audio is always generated, leaving sound out of the prompt does not give you silence - it gives you whatever the model picks. Name the ambience you want, or say "quiet room tone".

API quick start

# Generate the idempotency key once; reuse it when retrying after a lost response.
export SPICY_IDEMPOTENCY_KEY="${SPICY_IDEMPOTENCY_KEY:-$(uuidgen)}"
curl -X POST https://api.spicyapi.ai/api/v1/jobs/createTask \
  -H "Authorization: Bearer $SPICY_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $SPICY_IDEMPOTENCY_KEY" \
  -d '{
  "model": "kling/4.0/text-to-video",
  "input": {
    "prompt": "A slow aerial shot drifting over a misty pine forest at sunrise",
    "resolution": "720p",
    "aspect_ratio": "16:9",
    "duration_seconds": 5
  }
}'

# Copy data.taskId from the response into TASK_ID, then poll the status.
TASK_ID="TASK_ID_FROM_RESPONSE"
curl "https://api.spicyapi.ai/api/v1/jobs/recordInfo?taskId=$TASK_ID" \
  -H "Authorization: Bearer $SPICY_API_KEY"

See the API reference for authentication, the complete schema, polling, and error handling.

Missing the model you need?

Tell us what to add and where you would use it. Requests go to the people who decide what gets onboarded next.

KLING'S 30-SECOND MODEL

Kling 4.0 API at a glance

Up to 30 seconds per clip, 21:9 ultrawide, keyframes from first to last, and references from images, video and voice - with sound on every clip.

Longest clip: 30S; Any whole number of seconds from 3 to 30 on every endpoint.
Ultrawide: 21:9; Plus 16:9, 9:16 and 1:1, at 720p or 1080p.
Frames to steer: 10; Opening frame, closing frame and up to 8 in-between keyframes.
Reference images: 10; Plus up to 5 reference videos and 3 voice clips, each named in the prompt.
Sound included: STEREO; Every clip carries a two-channel soundtrack at no extra charge.

Results

What you can make with Kling 4.0

From text
A long take with its own sound
One paragraph gave us the brass wheel turning, the dome sliding open on the stars and a whispered line, with the motor rumble and wind on the soundtrack.
Text to video, 5 seconds at 720p, 16:9, with the dome rumble and wind on the soundtrack.
From frames
Pin the start and the end
An opening frame of her at the telescope and a closing frame of her leaning on the railing became one continuous move between the two.
Image to video with an opening and a closing frame, 5 seconds at 720p.
From references
Cast a scene from photos
A studio portrait as @Image1 and a photo of a chrome trailer as @Image2 became her stepping out of that trailer at blue hour.
Reference to video from two images (@Image1 the woman, @Image2 the trailer), 5 seconds at 720p.

More clips made with Kling 4.0

Image to video, 5 seconds at 720p, aspect ratio following the opening frame.
5s
Text to video, 5 seconds at 720p, 9:16 vertical.
5s
Reference to video from two images (@Image1 the man, @Image2 the motorcycle), 5 seconds at 720p.
5s

Content test

Is Kling 4.0 uncensored? What we tested

We ran each level on our live setup and checked every clip frame by frame. A pass means the content was actually on screen, not just that the request went through.

Content levels tested on Kling 4.0, by mode
Level	Text to video
MildThe lightest prompts in the test	Not tested yet
ModerateWhere filtered models start to refuse	RefusedTested Oct 16, 2026
StrongThe middle of the scale	Not tested yet
Very strongWhere most models start to hold back	RefusedTested Oct 16, 2026
MaximumThe top of the scale	Not tested yet

Renders: on screen as written. Softened: generated, but milder than what was written. Refused: rejected. Not tested yet: we have not run that level on that mode.

How each mode behaves

All modes: Prompts and inputs are screened before generation. Requests from the Moderate level up are refused within a few seconds and not charged; Mild-level wardrobe is not reliable either. Every clip carries generated audio - there is no silent option except keeping a reference video's own sound.
All modes: Generation is slow: 3-second clips took about 1.5 to 4 minutes in our tests, and requests on the same endpoint run one at a time, so a batch queues up.
Image to video: Up to 8 in-between keyframes. The opening, in-between and closing frames should show the same character and set from the same camera position, or the clip has to cut or morph to reach them.
Reference to video: Reference videos may add up to 30 seconds in total and their seconds are billed too, each clip rounded up to a whole second; audio clips must be 5 to 30 seconds each.

Use cases

What to build with the Kling 4.0 API

Long establishing shots
Up to 30 seconds lets a camera move play out in full - a slow crane, a drive down a road, a reveal - without stitching clips.
Storyboarded motion
Upload the key poses as opening, in-between and closing frames and let the model fill the motion between them.
Consistent characters and props
Reference a face, an outfit and a vehicle once and cast them into new scenes, with a voice sample if the character speaks.

Prompts

Kling 4.0 prompt examples you can copy

Paste one into the playground as it is, then change one thing at a time.

Reference to video
Cast list first, then the shot
Result from this prompt
@Image1 is the woman, @Image2 is the chrome trailer. At blue hour in the high desert, the woman from @Image1 steps out of the door of the trailer from @Image2 holding two steaming enamel mugs; the string lights along its awning blink on one by one and their glow slides across the polished aluminium. She stops, looks up at the first stars and smiles, then glances at the camera. Slow push-in. Sound: the creak of the trailer door, boots on cracked earth, a soft desert wind, a distant owl.
Why it works. Opening with one sentence per reference tells the model which face and which object to keep; the rest of the prompt only has to describe the action, the camera and the sound.
Try this prompt
Text to video
Action, camera, then sound
Result from this prompt
Inside a white observatory dome on a high-desert mesa at indigo dusk. An astronomer in her late thirties with short silver-streaked hair, wearing a charcoal wool sweater and fingerless gloves, turns a heavy brass wheel; with a deep rumble the dome slit slides open and a band of deep-blue sky full of first stars pours light across the polished chrome telescope. She looks up, breath visible in the cold air, and says quietly, "There you are." Slow crane up from her hands on the wheel to the opening sky. Sound: the grinding rumble of the dome motor, desert wind, the soft hum of equipment.
Why it works. The prompt runs in the order things happen and keeps the sound on its own line, so the rumble, the wind and the spoken line all land where they should.
Try this prompt
Image to video
Pin the ending with a second frame
Result from this prompt
She takes one last look through the open slit of the dome, then turns around, leans back against the steel railing and crosses her arms with a small satisfied smile as more stars come out over the mesa. Static wide shot, her braid and coat stirred by the wind. Sound: steady desert wind, the faint whirr of the telescope motor.
Why it works. With a closing frame set, the prompt only needs to describe how she gets from one pose to the other.
This mode starts from your own image: add it in the playground before you run.
Try this prompt

Compare

Kling 4.0 vs similar models: which one to pick

Model	Modes	From	Content	Pick it when
Kling 4.0This page	Image to video · Reference to video · Text to video	$0.216 / output second	Filtered — This model refuses requests beyond its own limits outright. Measured, not assumed — models we have not measured carry no label at all.	You want clips longer than 15 seconds, 21:9 ultrawide, in-between keyframes, or references from videos and voice clips as well as images.
Kling 3.0	Character animation · Image to video · Text to video	$0.067 / output second	Filtered — This model refuses requests beyond its own limits outright. Measured, not assumed — models we have not measured carry no label at all.	You need 4K, an audio switch, or Motion Control to copy a performance onto your own character.
Kling O3	Image to video · Reference to video · Text to video · Video editing	$0.067 / output second	Filtered — This model refuses requests beyond its own limits outright. Measured, not assumed — models we have not measured carry no label at all.	You want to edit a video you already have, or reference photos at up to 4K.
Kling 3.0 Turbo	Image to video · Text to video	$0.112 / output second	Filtered — This model refuses requests beyond its own limits outright. Measured, not assumed — models we have not measured carry no label at all.	You want a faster, cheaper clip with sound and do not need more than 15 seconds.

Pricing

Kling 4.0 API pricing

A 5-second clip starts at $1.08 at 720p and $1.62 at 1080p.

The playground shows the exact price for your settings before you run. Failed or timed-out tasks are usually not charged. Some models charge for content refusals as stated on their model page; check those billing terms before submitting.

See every price tier

Get started

How to use the Kling 4.0 API in three steps

Step 1
Draft short at 720p
Find the framing and the motion with 3 to 5 seconds at 720p before paying for a long or 1080p render.
Step 2
Write a Sound: line
Every clip has audio. Name the ambience and any spoken line in quotes so you get the sound you want.
Step 3
Name your references
On reference to video, open with "@Image1 is …, @Image2 is …" before describing the shot.

Open the playground Get an API key

Parameter reference and code samples are in the API and Readme tabs at the top of this page.

Why SpicyAPI

Why run Kling 4.0 on SpicyAPI

No filter from us
We add no content filter of our own. What the model renders is what you get back.
Tested, not guessed
Content labels come from real generations we ran and checked, with the test date on the page.
Pay per result, in dollars
No subscription and no credits. Failed or timed-out tasks are usually not charged. Some models charge for content refusals as stated on their model page; check those billing terms before submitting.
One key for every model
One API key and balance cover image, video, audio and chat models, so you can explore more models from the same account.

FAQ

Kling 4.0 API FAQ

About Kling 4.0

Kling 4.0 brings Kling's newest video model to one API: text to video, image to video with opening, closing and in-between frames, and reference to video from images, videos and voice clips. Clips run 3 to 30 seconds at 720p or 1080p, in 16:9, 9:16, 1:1 or 21:9, with a stereo soundtrack on every clip. Requests are screened and anything from the Moderate level up is refused without a charge. SpicyAPI bills per second of output in US dollars, with no subscription.