HiAPI
Blog

Guide

Kling 4.0 API: Release Date, Announced Specs and What to Use Now

Kling 4.0 was announced on September 28, 2026 with no public API yet. What was announced, when to expect it, and how to build on Kling 3.0 today.

Oct 15, 2026 4
可灵 4.0 API 文章封面图:深墨绿底上两组宽扁铜色与暗绿色缎带交织成扁平网格,无文字。
Announced
Sep 28, 2026
Max length (announced)
30 s
Kling 3.0 from
$0.10/s

Short answer: there is no public Kling 4.0 API yet. Kuaishou announced Kling 4.0 on September 28, 2026, and press coverage says the full model is scheduled for October 2026 without a specific date. So far only a limited early-access "Flash" version is live inside Kling's own app, with no API and no published pricing. On HiAPI, Kling 4.0 has a model page marked Coming Soon, where you can sign up to be notified when it goes live.

Until then, Kling 3.0 is the version you can call, and the code below is written so that switching to 4.0 later means changing one model name.

What has been announced

According to reports on the announcement (SlopTV, GuruFocus):

Item Announced Available today
Single generation length Up to 30 seconds Flash: 3–20 seconds
Keyframes Up to 10 Not yet
Multimodal references Up to 15 per request Not yet
Resolution Up to 4K Flash: up to 720p
Release Full model in October 2026, no day announced Flash in limited early access since September 28
API and pricing Not published None

These are announced figures, not tested ones. We'll update this page when Kling 4.0 is callable on HiAPI.

What you can use now: Kling 3.0

HiAPI runs three Kling 3.0 lines through the same POST /v1/tasks endpoint:

Model Endpoints 720p / second 1080p / second 4K / second
Kling 3.0 kling-3.0/* text-to-video, image-to-video (first or first + last frame) $0.10, $0.143 with sound $0.129, $0.193 with sound $0.479
Kling 3.0 Omni kling-3.0-omni/* text, image, reference-to-video, video transformation $0.10, $0.129 with sound $0.129, $0.1645 with sound $0.479
Kling 3.0 Turbo kling-3.0-turbo/* text-to-video, image-to-video $0.13 $0.16 —

Live HiAPI prices, checked October 15, 2026. Failed tasks aren't charged.

A Kling 3.0 clip from the model page, 5 seconds:

A lone red fox sprints across a windswept snowy ridge at golden hour, powder snow kicking up behind each stride, then leaps over a frozen creek in a graceful arc. Cinematic, dramatic side lighting, ultra-detailed fur, shallow depth of field, smooth motion.

Generate with this prompt → (5 seconds, 16:9)

kling-3.0/text-to-video model page example, 5 seconds. At 720p without sound this costs $0.50.

Write code you can switch to Kling 4.0 later

Keep the model name in configuration and the request shape stays the same. On HiAPI every video model uses the same task envelope, so moving to a new model is usually a config change plus whatever new input fields it adds.

import os, time, uuid, requests

API = "https://api.hiapi.ai/v1/tasks"
HEADERS = {"Authorization": f"Bearer {os.environ['HIAPI_KEY']}"}
KLING_MODEL = os.environ.get("KLING_MODEL", "kling-3.0/text-to-video")  # switch here later


def generate(prompt: str, **params) -> str:
    r = requests.post(API, json={"model": KLING_MODEL, "input": {"prompt": prompt, **params}},
                      headers={**HEADERS, "Idempotency-Key": str(uuid.uuid4())}, timeout=30)
    r.raise_for_status()
    task_id = r.json()["data"]["taskId"]
    while True:
        data = requests.get(f"{API}/{task_id}", headers=HEADERS, timeout=30).json()["data"]
        if data["status"] == "success":
            return data["output"][0]["url"]
        if data["status"] == "fail":
            raise RuntimeError(data.get("error"))
        time.sleep(10)


print(generate("A red fox sprints across a snowy ridge at golden hour",
               duration=5, resolution="720p", aspect_ratio="16:9", sound=False))

Full parameters, image-to-video, and error handling are in the Kling 3.0 API guide.

Which Kling 3.0 to pick while you wait

  • Clips with dialogue or sound effects, multi-shot prompts, or 4K: kling-3.0. See the Kling 3.0 prompt examples.
  • A consistent character or product from reference images, or following a reference video's camera: kling-3.0-omni.
  • One start frame, simple flat pricing, built-in sound is fine: kling-3.0-turbo.

FAQ

When will the Kling 4.0 API be available?

Kuaishou has said the full model launches in October 2026 but hasn't named a date or announced API access. HiAPI will open it once it is callable and priced; sign up on the model page to be notified.

How much will Kling 4.0 cost?

No price has been published. HiAPI shows prices on the model page only once a model is callable.

Can I use Kling 4.0 Flash through an API?

Not at the moment. The Flash version is limited early access inside Kling's own product.

Will my Kling 3.0 code work with Kling 4.0?

The task flow (submit, poll, download) will be the same on HiAPI. New capabilities like more keyframes or references will likely come with new input fields, so check the model's docs when it launches.

Start with Kling 3.0 today: generate the fox clip on the model page.

Related articles

View all articles →

Start building with persistent AI artifacts

Get your API key and call image, video and audio models through one task-based API.

Read the API docs