TL;DR
- Kling 4.0 (
kling-4.0/text-to-videoandkling-4.0/image-to-video) is billed per second on HiAPI: $0.22/s at 720p, $0.33/s at 1080p. Text-to-video and image-to-video cost the same. - Synchronized audio is included in every clip at no extra charge. A 5-second 720p clip is $1.10, and $10 covers about 45 seconds of video.
- It is live on HiAPI: open the model page to try it, and new accounts get 200 free credits.
Kling 4.0 price per second
| Resolution | Per second | 5 s | 10 s | 30 s (max) |
|---|---|---|---|---|
| 720p (default) | $0.22 | $1.10 | $2.20 | $6.60 |
| 1080p | $0.33 | $1.65 | $3.30 | $9.90 |
HiAPI list price, checked October 26, 2026.
Only two things change the price: resolution and seconds. Any whole number from 3 to 30 seconds is accepted, with no minimum charge. Aspect ratio and the first-frame image do not affect it. In credits, 720p is 440 credits per second, so a 5-second clip is 2,200 credits. Failed tasks are not charged.
Same prompt at 720p and 1080p
An old steam train pulls slowly out of a mountain station at dawn, thick white steam billowing into low golden sunlight, pine forest on both sides, slow lateral tracking shot, cinematic depth of field, natural engine rumble and distant birdsong
| 720p, 5 s, $1.10 | 1080p, 5 s, $1.65 |
|---|---|
Generate with this prompt → (pick 16:9 and 5 seconds on the model page, then the resolution)
kling-4.0/text-to-video, 16:9, 5 seconds. Both clips have a synchronized audio track.
What you can see:
- The 1080p file is 1920×1080 and the 720p file is 1280×720. If the clip goes into a 1080p edit or onto a big screen, render 1080p rather than upscaling.
- The framing differs: the 720p clip tracks along the rails through the forest, while the 1080p clip is a wide side view of the station with the sun behind the steam. The same prompt gives a new take every run, so lock the shot at 720p and render the final at 1080p.
- Both clips have a synchronized audio track. The 1080p clip took about 4 minutes from submit to done.
How to spend less
- Iterate at 720p, render the final at 1080p. The prompt stays the same; only
resolutionchanges. That saves $0.55 per 5-second draft. - Set
durationper shot. It defaults to 5 seconds and cost scales linearly, so a 30-second 1080p clip is $9.90. - Direct the audio in the prompt. Every Kling 4.0 clip has sound and there is no off switch. Describe the sound you want and replace it in the edit if needed.
- Validate before you submit. A bad resolution or duration returns HTTP 400 right away. No task is created and nothing is billed.
Kling 4.0 vs Kling 3.0 vs Seedance 2.5 pricing
| Model | 720p / s | 1080p / s | Clip length | Audio |
|---|---|---|---|---|
| Kling 4.0 | $0.22 | $0.33 | 3–30 s | Always on |
| Kling 3.0 | $0.10, $0.143 with sound | $0.129, $0.193 with sound | 3–15 s | Optional |
| Kling 3.0 Turbo | $0.13 | $0.16 | 3–15 s | Always on |
| Seedance 2.5 | $0.204 | $0.516 | 4–30 s | — |
- Single shots longer than 15 seconds: Kling 4.0 and Seedance 2.5 both go to 30 seconds. At 1080p, Kling 4.0 costs about 36% less per second.
- Short clips on a budget: Kling 3.0 with sound is $0.715 for 5 seconds at 720p, against $1.10 on Kling 4.0.
- 4K output: use Kling 3.0 ($0.479/s). Kling 4.0 tops out at 1080p.
For side-by-side clips from the two generations, see Kling 4.0 vs Kling 3.0.
Monthly budget (5-second clips)
| Clips per month | 720p | 1080p |
|---|---|---|
| 100 | $110 | $165 |
| 1,000 | $1,100 | $1,650 |
| 10,000 | $11,000 | $16,500 |
FAQ
Does audio cost extra?
No. Every Kling 4.0 clip includes synchronized audio, and the price depends only on resolution and seconds.
Is image-to-video priced differently?
No. Both endpoints are $0.22/s at 720p and $0.33/s at 1080p. The first-frame image is not billed separately.
Is there a 4K tier?
No. Kling 4.0 offers 720p and 1080p. For 4K, use Kling 3.0.
Are failed tasks charged?
No. Invalid parameters are rejected with a 400 before a task exists, and failed generations are not billed.
The request code is in the Kling 4.0 API guide, and ready-made prompts are in Kling 4.0 prompts. Or copy the steam train prompt above and generate your first clip.
