Kuaishou video model · 3.0 generation · 720p and 1080p only

Kling 3.0 Turbo Video Generator

Run 3–15 second clips at 720P or 1080P on the value tier of Kuaishou’s Kling 3.0 generation, where multi-shot and native audio are unchanged and image reference is absent entirely.

Try Kling 3.0 Turbo
01

Create with Kling 3.0 Turbo

Pick text or a single first frame, set duration, resolution, and aspect ratio, then read the point estimate before you submit.

LIVE
5s
2s15s

This model does not support reference video input

Results

Creation history

Generated content will appear here
02 / Overview

Kling 3.0 Turbo specifications

Facts reviewed:

Developer
Kuaishou / Kling AI
Official model ID
kling-3.0-turbo
Generation and tier
Kling 3.0 generation · value tier
Official positioning
Consistently excellent results, with even better value
Official resolution
720P and 1080P; no 4K row on the price table
Official clip length
3–15 seconds
Official aspect ratio
16:9, 9:16, 1:1 on the text-to-video endpoint
Official API price
0.8 Units (USD 0.112) per second at 720P; 1.0 Unit (USD 0.14) at 1080P
Official release date
17 June 2026
Platform route
kling-3.0-turbo

Documented model abilities next to this page’s controls

The official capability map describes the model. This page runs one platform route, and the two are kept separate here because the gaps are the whole reason this tier is priced where it is.

OFFICIAL

Documented for the official model

  • Text-to-video and image-to-video, both reachable at 720P and 1080P
  • Single-shot generation, multi-shot generation, and native audio, each marked supported on the same rows as Kling 3.0 and Kling 3.0 Omni
  • Multi-shot written in the official fixed format, up to six shots, each at least one second, with shot durations summing to the clip length
  • First-frame image-to-video only, with the official note that first-and-last-frame and last-frame-only are not supported yet
  • Avatar, lip sync, video to audio, and text to audio, which the official map marks as applying to all model versions rather than to a tier
  • 3–15 seconds, with the aspect ratio parameter present on the text-to-video endpoint only
  • No 4K anywhere in this tier, and no element control, video reference, or multi-image input
LIVE PLATFORM

Exposed by this workspace today

  • Text-to-video and image-to-video with one image; first-and-last-frame is not exposed
  • A 2–15 second selector, against an official 3–15 second range
  • A resolution selector listing 720p and 1080p, with 720p as the default
  • Five aspect ratios, including 4:3 and 3:4, which the official text-to-video parameter does not list
  • No sound switch, matching the official note that this tier has no audio parameter to toggle
  • Points quoted before submission on the live estimate

Two differences are worth reading twice. The duration selector starts at 2 seconds while the official capability map and both API documents start at 3, so the shortest clip this page offers is shorter than the shortest clip the model documents. The aspect-ratio selector also lists 4:3 and 3:4, which the official text-to-video endpoint does not enumerate. Both are platform controls rather than documented model capabilities, and the price you see quoted here is a platform point estimate, not the official Units figure.

Model brief

Where Kling 3.0 Turbo sits in the Kling 3.0 generation

Kling 3.0 Turbo is the value tier of Kuaishou’s Kling 3.0 generation. The official capability map registers it as kling-3.0-turbo alongside kling-v3 and kling-v3-omni, and the official update log dates its release to 17 June 2026. Kuaishou sums the tier up in one line as consistently excellent results with even better value, and the official API price table holds that description up rather than contradicting it.

That price table is the first place this tier separates itself. At 720P it is billed at 0.8 Units per second, which the same document converts to USD 0.112, and at 1080P at 1.0 Unit, or USD 0.14. The standard Kling 3.0 model carries 0.9 and 1.2 Units per second on the equivalent native-audio rows. The cheaper tier is genuinely cheaper, and both numbers come from one official table rather than from a reseller’s listing.

What the tier keeps matters as much as what it costs. The capability map runs the same rows down every model in the generation, and on single-shot generation, multi-shot generation, and native audio, Turbo is marked supported in exactly the same way as Kling 3.0 and Kling 3.0 Omni. Anyone who needs several shots in one pass and sound generated with the picture is not giving up either of them by choosing this tier.

The concessions are narrower and more specific than the word Turbo tends to imply. Turbo is the only model in the 3.0 generation with no image reference of any kind. First-and-last-frame is marked unsupported, element control is unsupported, video reference is unsupported, and multi-image input is unsupported. What remains is a single first frame, which is why the image-to-video endpoint ships with a reusable image instead of a timeline of references.

Resolution is the second concession and it is absolute. The price table shows 720P and 1080P for this tier and a dash where the 4K column would be, while Kling 3.0 and Kling 3.0 Omni both carry 3.0 Units per second at 4K. This is not a marketing omission that a different endpoint fills in; it is the same document declining to price a resolution for this model.

This page runs the platform route kling-3.0-turbo rather than a direct account on Kuaishou’s own API. Two deviations follow from that and are stated separately throughout. The workspace exposes 2 to 15 seconds against an official 3 to 15 seconds, and its aspect-ratio selector lists 4:3 and 3:4, which the official text-to-video parameter does not offer. Everything else on this page — resolution, audio behaviour, shot count, per-second Units — is reported the way the official documents report it.

03 / Capabilities

What the Turbo tier keeps, and what it gives up

Turbo is worth choosing when the per-second price is the constraint and a single starting frame is the whole input.

01

The lowest per-second rate in the generation

At 0.8 Units per second for 720P and 1.0 Unit for 1080P, this tier sits below the standard Kling 3.0 model at 0.9 and 1.2 Units on the same native-audio rows, and the gap comes from a single official table.

02

Text to video

Describe subject, action, place, light, camera movement, and sound in one instruction and get a usable clip back without supplying any image at all.

03

Image to video from one frame

Animate a single still while holding its composition, which is the only reference input this tier documents.

04

Multi-shot generation as a documented capability

The capability map marks multi-shot supported on Turbo on the same row as the other two 3.0 models, so several consistent shots in one pass are not reserved for a higher tier.

05

Native audio on the same row as the standard model

Sound is marked supported for this tier exactly as it is for Kling 3.0 and Kling 3.0 Omni, and the price table prices Turbo on a native-audio basis rather than offering it as an upgrade.

06

Six shots addressed in one prompt

The official format writes each shot as a numbered index, a duration in seconds, and the shot description, separated by semicolons, with up to six shots and each one at least a second long.

07

A starting point at three seconds

The documented range begins at 3 seconds rather than at a fixed five, so a short beat can be generated on its own instead of being cut out of a longer clip.

08

No 4K, stated plainly

Resolution stops at 1080P, so a delivery specification that names 4K sends the approved direction to a tier where the price table does list 3.0 Units per second for it.

04 / Scene ideas

When Kling 3.0 Turbo is the right pick

Turbo is worth choosing when the per-second price is the constraint and a single starting frame is the whole input.

High-volume first passes

Push many prompt directions through an initial round at the lowest per-second rate in the generation, then take only the survivors further.

Multi-shot sequences that need sound

Build a short sequence with several shots and generated audio in a single pass, because both are documented for this tier rather than withheld from it.

Vertical and square social cuts

Produce 9:16 and 1:1 variants for the same idea without paying a higher per-second rate for each one.

Prompt and pacing comparisons

Change one variable at a time across a slow push, a locked-off frame, and a steady lateral track to see which reading the scene supports.

Per-second budgeting on client work

Quote from a published per-second figure at 720P or 1080P instead of from a subscription tier whose video allowance has to be reverse-engineered.

Drafts for a 4K finish

Lock the shot count, timing, and sound design on the cheap tier, then re-run only the approved direction where 4K is available.

05 / Prompt lab

Prompts suited to the Turbo tier

These use the official multi-shot format, writing each shot as an index, a duration, and a description separated by semicolons. One camera movement per shot, and no timecodes inside the sentences.

SCENE 01text to video

Two-shot product insert at 720P

Use the cheapest resolution for a lighting and reflection test where the framing is already decided.

View prompt
shot 1, 3, a brushed steel coffee grinder on a slate counter, hard side light from the left, dust visible in the air; shot 2, 4, slow push in as a hand turns the crank once and grounds fall into a glass jar, low mechanical clicks. Keep the grinder shape, the slate grain, and the jar level unchanged. Restrained industrial look, no logos, no captions.
SCENE 02text to video

Three-shot street sequence with native audio

Check that shot continuity and sound stay together across cuts at the value tier.

View prompt
shot 1, 3, wide view of a rain-slick street at dusk, neon signs reflecting on the asphalt, steady locked-off camera; shot 2, 4, a courier in a wet jacket walks from left to right through the frame, water splashing at each step; shot 3, 3, close on wet boots stepping into a shallow puddle as the sign above flicks off. Keep colour and face consistent across shots, cool evening light, realistic rain, no subtitles.
SCENE 03image to video

Single-frame portrait move

Confirm that one supplied frame survives a small move and a natural performance before a longer pass.

View prompt
Preserve the uploaded person’s identity, hairstyle, wardrobe, lighting direction, and background layout. Add one natural breath, a slow turn of the head toward camera, and a gentle five-percent push in. Keep facial proportions stable and skin texture realistic. Quiet room ambience, no added objects, no text.
SCENE 04text to video

Vertical six-second social beat

Use the documented starting point of three seconds per shot to keep a vertical cut tight.

View prompt
shot 1, 3, a ceramic cup of black coffee on a windowsill, morning light raking across the rim, static camera; shot 2, 3, steam rises and bends to the right as a hand enters frame and lifts the cup, ceramic scrape on wood. Keep the cup shape and the window frame unchanged, warm morning light, kitchen room tone, no text.
SCENE 05text to video

Material and geometry stress test

Isolate one camera movement to see how the tier renders a hard surface and its reflections.

View prompt
One continuous shot, no cuts. A dark marble countertop with visible veining, locked-off camera at counter height. A single drop of water falls and spreads slowly across the surface while the reflection of an overhead strip light stretches and shortens. Keep the marble pattern and the light position unchanged for the whole clip, quiet room tone, no hands, no other camera movement, no text.
COMPARE

Kling 3.0 Turbo, Kling 3.0 and Kling 3.0 Omni on one official table

Every row below is read from one official capability map and one official API price table, both of which list all three models in the same document. Nothing here is assembled from a third-party comparison.

Kling 3.0 Turbo, Kling 3.0 and Kling 3.0 Omni on one official table
ComparisonKling 3.0 TurboKling 3.0Kling 3.0 Omni
Official model IDkling-3.0-turbokling-v3kling-v3-omni
Official resolution720P, 1080P720P, 1080P, 4K720P, 1080P, 4K
Official clip length3–15 seconds3–15 seconds3–15 seconds
Single-shot, text to videoSupportedSupportedSupported
Multi-shot, text to videoSupportedSupportedSupported
Native audio, text to videoSupportedSupportedSupported
First and last frame, image to videoNot supportedSupportedSupported
Element control, image to videoNot supportedSupportedSupported
Video reference, image to videoNot supportedNot supportedSupported
Multi-image inputNot supportedNot supportedSupported
API price at 720P, per second0.8 Units (USD 0.112)0.9 Units (USD 0.126)0.8 Units (USD 0.112)
API price at 1080P, per second1.0 Unit (USD 0.14)1.2 Units (USD 0.168)1.0 Unit (USD 0.14)
API price at 4K, per secondNot offered3.0 Units (USD 0.42)3.0 Units (USD 0.42)
Official release date17 June 202625 February 202625 February 2026
LIVE PRICING

How Kling 3.0 Turbo is billed

Kuaishou bills this tier by the second of generated video at 0.8 Units (USD 0.112) for 720P and 1.0 Unit (USD 0.14) for 1080P, with native audio included in both figures rather than charged as an extra. The same table lists the standard Kling 3.0 model at 0.9 and 1.2 Units on the equivalent rows and marks the 4K row as not offered for Turbo. The published update log dates the release to 17 June 2026. This site charges platform points instead, so the live estimate in the workspace is the price that applies here.

Open the workspace to see the current estimate
01

Platform points

The workspace quotes points before you submit. Points are set by this site and are not the official Units figure.

02

Resolution and duration

The official tier documents 720P and 1080P from 3 to 15 seconds. This workspace exposes 2 to 15 seconds and lists 4:3 and 3:4 among its aspect ratios, neither of which has an official counterpart for this model.

06 / Workflow

Running a Kling 3.0 Turbo generation end to end

  1. 01

    Start from text or one frame

    Text-to-video and single-image image-to-video are the two modes this route exposes. First-frame-plus-last-frame is an official capability of other models in the generation but not of this one, and not a control here either.

  2. 02

    Write the shots in the official format

    Number each shot, give it a duration in seconds, and describe it, separating shots with semicolons. Up to six shots, each at least one second, with the durations adding up to the total clip length.

  3. 03

    Keep one camera movement per shot

    A push, a pull, a lateral track, or a static frame. Two movements in one sentence split the model’s attention and usually produce neither.

  4. 04

    Hold subject definitions steady across shots

    Give each subject two or three fixed static traits and reuse the same wording in every shot so a multi-shot pass does not blur two characters into one.

  5. 05

    Judge the draft at 720P and promote only what holds

    Check identity, geometry, on-screen text, motion, and audio on the cheap resolution, then move an approved direction to a tier that documents 4K if the delivery specification demands it.

Review before publishing

Turbo tier boundaries, billing, and disclosure

Lower cost does not remove the usual failure modes. Identity drift across shots, unstable geometry, unintended cuts, inaccurate on-screen text, and audio that drifts from the picture can all appear at any price tier. Every draft is review material, and the full clip plus its complete soundtrack should be watched before publication rather than judged from a few frames.

Two service boundaries on this page are not interchangeable, and both come from the difference between a platform control and a documented model capability. The official capability map and both API documents give this tier a 3 to 15 second range, while the duration selector here starts at 2 seconds, so the shortest clip this page can produce is shorter than any clip the model documents. The aspect-ratio selector adds 4:3 and 3:4, which the official text-to-video parameter does not enumerate. Neither is a defect in the model; both are reasons to read the selector as a platform control.

The absence of image reference is the boundary most likely to be missed. A reader who is used to the reference-driven workflows of other Kling models may expect element control or a video reference on this page and will not find either, because the official capability map marks both unsupported for this tier. A single first frame is the whole image input, and multi-image input is unsupported as well. Work that depends on holding a character or a product across shots from supplied material needs a model whose table marks those rows supported.

Price and availability statements here are dated. The per-second figures come from the official API price table read on 13 September 2026, and the release date comes from the official update log. That same log carries a retirement notice for older Kling models taking effect on 15 September 2026, which does not touch this tier but is close enough to this page’s review date to warrant re-checking rather than assuming. Upload only material you are authorised to use, and do not impersonate people, mislead viewers, or omit disclosures that apply to synthetic media.

Sources

Kling documents behind this page

Official model identity, capabilities, limits, and pricing are separated from this platform’s controls, points, and availability. Where the two differ, the difference is stated rather than smoothed over.

07 / FAQ

Kling 3.0 Turbo questions, answered

Is Kling 3.0 Turbo a lower-quality version of Kling 3.0?
The official documentation does not frame it that way. Kuaishou describes the tier as consistently excellent results with even better value, and the capability map marks single-shot generation, multi-shot generation, and native audio as supported on Turbo on the same rows as Kling 3.0. What the tier gives up is image reference control and the 4K row. Third-party pages that present Turbo as a speed-or-quality trade are not quoting an official comparison.
Does Kling 3.0 Turbo support 4K?
No. The official API price table lists 720P and 1080P for this model and shows a dash in the 4K column, while Kling 3.0 and Kling 3.0 Omni both carry 3.0 Units per second there. This page does not list 4K in any control for that reason.
How much does Kling 3.0 Turbo cost?
Kuaishou bills it by the second: 0.8 Units per second at 720P, which the same table converts to USD 0.112, and 1.0 Unit per second at 1080P, or USD 0.14. Both figures include native audio. The standard Kling 3.0 model is listed at 0.9 and 1.2 Units per second on the equivalent rows.
What is the official model ID?
kling-3.0-turbo. It is the identifier on the official capability map and on both API endpoints, POST /text-to-video/kling-3.0-turbo and POST /image-to-video/kling-3.0-turbo. This platform uses the same identifier as its route name, so there is no translation layer between the page and the model.
How do I write a multi-shot prompt?
The official format is a numbered shot index, a duration in seconds, and the shot description, with shots separated by semicolons: shot 1, 3, description; shot 2, 4, description. Up to six shots, each at least one second, and the shot durations must add up to the total clip length. This model has no separate multi-shot parameter, so the format in the prompt is what produces multiple shots.
Can I upload several reference images?
No. Multi-image input is marked unsupported for this tier, as are element control and video reference. The image-to-video endpoint takes a single first frame, with the official note that first-and-last-frame and last-frame-only are not supported yet. If your work depends on several references, a model whose capability map marks those rows supported is the right route.
Does Kling 3.0 Turbo generate audio?
Native audio is marked supported on this tier on the same capability-map row as Kling 3.0 and Kling 3.0 Omni, and the price table charges Turbo on a native-audio basis rather than as an add-on. This workspace does not expose a sound switch on this route, which matches the official position that the tier has no audio parameter to toggle.
Is lip sync a Turbo-specific feature?
No. The official capability map lists Avatar, lip sync, video to audio, and text to audio on one line marked as applying to all model versions, which means the row does not distinguish between tiers. This platform does not expose those controls, so they are documented model abilities rather than buttons here.
What aspect ratios and clip lengths are supported?
The official text-to-video endpoint enumerates 16:9, 9:16, and 1:1 with 16:9 as the default, and the documented clip range is 3 to 15 seconds. The image-to-video endpoint has no aspect-ratio parameter at all. This workspace exposes 2 to 15 seconds and also lists 4:3 and 3:4, which have no official counterpart for this model.
Can I get a 2-second clip?
On this platform, yes, because the duration selector starts at 2 seconds. Officially, no: the capability map and both API documents give this tier a range that starts at 3 seconds. Treat 3 seconds as the model floor and 2 seconds as a platform control, and expect the shorter value to be an extension of what the model documents rather than something the official range covers.
How does turbo speed compare with Kling 3.0?
The reachable official pages do not publish a speed comparison or a generation-time figure for either model, so this page makes no claim about how much faster the tier is. The official release entry for 17 June 2026 describes the parameter optimisation as reducing parameter conflicts and making integration easier, which is a statement about the interface rather than about rendering time.
Are third-party credit prices the same as the official price?
No. Several reseller pages publish their own credit figures per second for this tier, and those numbers are not the official Units figures and are not interchangeable with them. This page cites only the official API price table, and the point estimate shown in the workspace is this platform’s own price rather than either of the above.

Other Kling routes that take a different input

01

Kling O3

Compare the reference-driven tier, the only Kling model whose official rows mark video reference and multi-image input supported.

02

Seedance 2.5

Move to a different family when 1080p at 10-bit or longer clips matter more than the per-second rate.

03

Vidu Q3 Turbo

Compare another speed-oriented tier from a different provider on the same workspace.

04

Text to Video

Open the general prompt-led video workspace with model selection.

Render your next scene

Start from a prompt or a single frame

Return to the live workspace, write your shots in the official format, and review the current point estimate before generating.

Open Kling 3.0 Turbo workspace