Kling 3.0 Turbo API
Kling 3.0 Turbo 高速视频生成模型，支持 3～15 秒文生和单首帧图生视频、最多 6 个分镜，输出 720p 或 1080p。
What is Kling 3.0 Turbo
Kling 3.0 Turbo is the speed tier of Kuaishou's Kling 3.0 family, released on 2026-06-17. Each run produces a 3–15 second clip (default 5 s) at 720p or 1080p, in 16:9, 9:16 or 1:1 (default 16:9). Upstream materials claim up to 6 shots per generation plus native audio with lip-sync in five languages; these parameters are not listed in Roko's catalog, so treat the task response as authoritative.
Roko integrates the model through the /v1/videos/tasks async task API with a single video-task endpoint and two routes: text-to-video maps to kling-v3-turbo-text-to-video, and image-to-video maps to kling-v3-turbo-image-to-video. The image route accepts 1 image and inherits its aspect ratio; input modalities are text and image.
Pricing follows the Roko production catalog: $0.1166 per second at 720p and $0.1463 per second at 1080p, billed by output seconds — estimated at submission and settled on actual usage after the task completes. Kuaishou's official list prices (720p ¥0.8/s, 1080p ¥1/s) differ and are not converted or cross-applied.
Model pricing
Pricing follows the current production configuration and the billing dimensions supported by each route.
| Route | Price | Billing rule |
|---|---|---|
| 文生 / 图生视频 | 720p $0.1166/sec1080p $0.1463/sec | 按输出视频秒数 |
Cost examples
Examples use Roko catalog rates: $0.1166/s at 720p and $0.1463/s at 1080p, billed by output seconds, estimated at submission and settled on actual usage.
5 s × $0.1166/s = $0.583
Default duration at default resolution — the workhorse for social clips and batch A/B assets.
5 s × $0.1463/s = $0.7315
Same duration at 1080p; the premium buys sharper detail for paid placements.
15 s × $0.1463/s = $2.1945
Max duration at max resolution, suited to narrative multi-shot cuts (multi-shot is an upstream claim).
Model information
Only provider-documented routes, capabilities, and limits implemented by this gateway are listed.
kling-v3-turbo-text-to-video
- Limits
- —
- Duration
- 3–15s
kling-v3-turbo-image-to-video
- Limits
- Images 1
- Duration
- 3–15s
Key capabilities
Speed-first, cost-light
Turbo is the speed-optimized tier of the Kling 3.0 family, built for rapid iteration and high-volume output. Roko bills by output second: a full 15-second 1080p clip costs $2.1945.
First-frame image-to-video
Feed 1 image as the opening frame and the output inherits its aspect ratio automatically — no separate framing step. Ideal for animating product shots and posters.
720p/1080p, 3–15 seconds
Default 720p, switchable to 1080p; both tiers are metered per second ($0.1166 and $0.1463). Duration defaults to 5 s and can stretch to 15 s.
Multi-shot and audio are upstream claims
Upstream materials claim up to 6 shots, native audio, and lip-sync in English, Chinese, Japanese, Korean and Spanish. Roko's catalog currently lists only the two t2v/i2v routes; available parameters follow the task response.
Main parameters
Parameters come from structured model configuration; unsupported or unconfigured fields are omitted.
modelstringSubmodel ID
Default:kling-v3-turbo-text-to-video / kling-v3-turbo-image-to-videopromptstringKling 3.0 Turbo 高速视频生成模型，支持 3～15 秒文生和单首帧图生视频、最多 6 个分镜，输出 720p 或 1080p。
Default:RequireddurationintegerOutput video seconds
Default:3–15squalityenumCurrent billing
Default:720p / 1080paspect_ratioenumCapability
Default:16:9 / 9:16 / 1:1image_urlsarrayImage to video: 1
Default:Varies by routeCapabilities and usage guidance
Key features
- Text to video · kling-v3-turbo-text-to-video
- Image to video · kling-v3-turbo-image-to-video
- 720p / 1080p output
- 3–15s Duration
Limits
- Image to video: Images 1
- Capability source: https://evolink.ai/zh/kling-3-0-turbo
Limits
- Roko currently exposes only the text-to-video and image-to-video routes through the video-task endpoint; other Kling 3.0 capabilities are not available.
- Image-to-video is first-frame only: the output aspect ratio inherits the input image and cannot be overridden.
- No 4K: upstream places 4K, motion brush, and start/end-frame control in Kling 3.0 Standard, not Turbo.
- A successful submission is not a successful generation: tasks can fail on content moderation or invalid parameters — the terminal state is authoritative.
- Output video links are time-limited; download and re-host your assets promptly.
How it works
- Sign up, create a Roko API key, and top up or confirm your available balance.
- Submit an async task to /v1/videos/tasks with model set to kling-3.0-turbo, choose text-to-video or image-to-video, and set duration, resolution and aspect ratio.
- The API returns a task ID plus a cost estimate based on current rates.
- Poll GET /v1/tasks/{task_id}: a successful submission does not guarantee a successful generation — wait for a terminal state.
- On success the fee is settled by actual output seconds; download and re-host the video link promptly before it expires.
Use cases
Ad-creative A/B at scale
Spin up multiple candidate videos from one script; Turbo's speed positioning and per-second billing keep multi-variant testing affordable.
Bring product photos to life
Route 1 product image through image-to-video for a first-frame-driven clip whose aspect ratio matches the placement automatically.
Talking-head and shoppable shorts
Upstream claims native audio and multi-language lip-sync suit voiceover formats; Roko's catalog lists no audio parameters, so verify via the task response.
How it compares
Comparisons below draw on the Roko catalog and upstream public materials.
vs kling-o3
Same Kuaishou Kling family: Turbo wins on speed and per-second price, while kling-o3 targets higher fidelity with 4K, audio, and editing tools. Pick Turbo for fast volume, kling-o3 for ceiling quality.
vs happy-horse-1.1
happy-horse-1.1 focuses on multi-reference consistency for characters and products across shots; Turbo trades that for faster turnaround and a lower per-second rate on batch assets.
vs hailuo-3
hailuo-3 accepts omni-modal reference inputs with a broader capability surface; Turbo sticks to two lightweight routes at a lower per-second price, stretching budgets further.
FAQ
Which endpoint serves kling-3.0-turbo on Roko?
Submit async tasks to /v1/videos/tasks with the video-task endpoint: text-to-video routes to kling-v3-turbo-text-to-video and image-to-video routes to kling-v3-turbo-image-to-video, with model set to kling-3.0-turbo.
How much does kling-3.0-turbo cost?
Roko catalog pricing is $0.1166/s at 720p and $0.1463/s at 1080p, billed by output seconds. Kuaishou's official prices (¥0.8/s and ¥1/s) differ and are not converted.
Which durations, resolutions and ratios are supported?
3–15 s (default 5 s), 720p/1080p (default 720p), and 16:9, 9:16, 1:1 (default 16:9). Image-to-video inherits the input image's ratio.
Are audio and multi-shot supported?
Upstream claims up to 6 shots, native audio, and lip-sync in five languages, but Roko's catalog lists only the two routes without multi-shot or audio parameters — available parameters follow the task response.
What are the image-to-video constraints?
First-frame only: 1 input image sets the opening frame, the output inherits its aspect ratio, and no other framing override is available.
Why isn't the video available immediately?
Roko uses an async task model: submissions queue for generation and you poll status until a terminal state, then fetch the video link.
Is kling-3.0-turbo free?
No. Generation is billed by output seconds — estimated at submission and settled on actual usage against your account balance.
Can I use the videos commercially?
You are responsible for verifying rights and compliance of generated content under applicable terms and local regulations; Roko provides the model access channel.
Sources
Upstream capability claims come from Kuaishou's public Kling materials; Roko-callable parameters, routes, and prices come from the current production catalog. Verified: 2026-10-09.
- Kuaishou: Q2 2026 results and Kling 3.0 Turbo Official company disclosure: Kling 3.0 Turbo launch, audio-visual synchronization, creative-efficiency positioning, and Kling 3.0 native 4K rollout
- Kling AI developer pricing Official developer rate card for Kling 3.0 Turbo resolution tiers and native-audio pricing
- HowAIWorks: Kling 3.0 family Upstream family background: kling-v3-turbo positioning, audio languages, aspect ratios
- Roko API video task docs Async submit, task query, and billing lifecycle on Roko
- Roko API model catalog Current routes, prices, and generation config
Models in this family
Explore other available variants from the same model family.
Kling 3.0 Turbo
Kling O3
View model →Other video models
Explore more models available through the unified asynchronous video task API.
