Kling.ai headlines Kling 4.0: what an API buyer can call today
kling.ai now shows 'All-New Kling 4.0', but fal still says full launch is October. Sume lists kling-3 (4-15 s, 720p/1080p) and no 4.0 id. Prices inside.

Yes, Kuaishou's own Kling page now puts Kling 4.0 up front, but a banner is not an API id. The kling.ai home page I read on 2026-10-08 carries the heading "All-New Kling 4.0" and the line "Audio-Visual Performance for Realistic Results". It gives no launch date, no price and no endpoint. fal's Kling 4.0 explainer, read the same day, says the full launch is set for October 2026 with no day announced, and that a faster Kling 4.0 Flash has been in limited early access since September 28.
On Sume the video catalog lists Kling 3.0 as kling-3 and a separate motion-control route. There is no Kling 4.0 id in the catalog on origin/main, so there is nothing to select today. Until a 4.0 id appears in the catalog, plan and budget on kling-3.
What each page says, read 2026-10-08
The table keeps the consumer page, the provider explainer and the Sume catalog apart, because they answer three different questions: is it announced, is it callable, and what does a call cost.
| Source | What it says | Callable on Sume? |
|---|---|---|
| kling.ai home page | Heading "All-New Kling 4.0"; no date, price or API terms; no mention of 4.0 Flash | No id in the catalog |
| fal.ai Kling 4.0 explainer | Full launch set for October 2026, no date; Flash in limited early access since September 28; up to 30 s, 3 s minimum, Flash capped at 20 s | No id in the catalog |
| fal.ai Kling 4.0 explainer | Full model 720p, 1080p, 4K; Flash 720p only; up to 10 keyframes; up to 15 references (10 images, 5 videos) | Not exposed |
Sume kling-3 | 4-15 s, 720p or 1080p, 16:9, 9:16 or 1:1, no reference_*_urls; $0.14 per second silent, $0.21 per second with sound | Yes |
Sume kling/3.0/motion-control | Output length follows the motion video, 1-30 s declared; $0.1575 per second | Yes |
Why the Sume catalog is the check, not the banner
The catalog lists each model's resolutions, durations and reference types, and GET /v1/videos/models returns it. A vendor announcing a model does not add an id by itself; the provider endpoint, price and limits have to be added to the catalog. The practical test is one GET.
import os
import requests
r = requests.get(
"https://api.sume.com/v1/videos/models",
headers={"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"},
timeout=30,
)
r.raise_for_status()
ids = [m["id"] for m in r.json()["data"]]
print([i for i in ids if i.startswith("kling")])Budget a Kling 4.0 project on kling-3 rates now
Sume bills the provider list price times 1.25, rounded up to the cent per job. At that rate a 15-second Kling 3.0 clip is $2.10 silent or $3.15 with sound, and 720p and 1080p cost the same. A 30-second piece is two 15-second jobs, $4.20 silent or $6.30 with sound, joined on a timeline. Kling 4.0's 30-second single generation, if and when it is listed, will change that plan; the 4.0 price is not published on the pages I read, since fal says endpoint names and pricing go up at launch.
- Keep the prompt, first frame and aspect ratio in your own config so swapping
kling-3for a 4.0 id later is a one-line change. - Do not design around 10 keyframes or 15 references yet;
kling-3takes a first frame and an optional end frame, and no reference URLs. - If you need a 30-second single take today, Seedance 2.5 and Wan 3.0 both list 30 seconds on Sume.
What this page does not claim
I did not find a Kling 4.0 price, a Kling 4.0 API doc or a release day on the pages above. The Kling home page heading is the only first-party statement I could read, and it is a marketing banner. Re-read the Kling page and the Sume catalog before you commit a launch date to a client.
Sources
Related posts
More in Models
- MAI-Transcribe-2: 60 languages, 17 more than 1.5, and the Sume hint
Microsoft lists 60 languages for MAI-Transcribe-2, 43 for MAI-Transcribe-1.5, so 17 new ones. Sume STT takes a language_code hint or auto-detects.
- MAI-Transcribe-2-Streaming runs in 4 regions, MAI-Voice in 14
Microsoft lists four regions for MAI-Transcribe-2-Streaming and 14 for MAI-Voice-2.1; three overlap. Sume's STT and TTS requests have no region field.
- MAI-Voice-2.1 languages: 23 listed, no Japanese or Arabic voice
Microsoft's MAI-Voice-2.1 voice table covers 23 languages and 28 locale codes, with no Japanese or Arabic row. How Sume's TTS language field handles both.
- MAI-Voice-2.1-Flash: 150 ms end-to-end and a 45-second audio limit
Microsoft's news post gives MAI-Voice-2.1-Flash 150 ms end-to-end latency and 45 s of audio; the Learn page gives no latency number. What each page supports.
Written by Sume