Kling 4.0 AI Video Generator

Kling 4.0 is Kuaishou’s video model for a new scene, a photo sequence, or a reference pack. On MixVio each video is 3–30 seconds at 720p or 1080p, and native stereo stays on. Kling 4.0 ↗

Text, keyframes, or reference to video. Text is a prompt only. Keyframes use one start photo, up to 8 middle frames, and an optional end frame, and follow the start photo. Adding a frame lengthens the video when the current duration cannot hold it; shortening the duration keeps every photo and pauses the frames past that limit. Reference to video takes up to 10 photos, 5 videos, 7 named elements, and 3 voice references, 15 items in total.

Check the file before you generate. Each photo is up to 30 MB, with the short side at least 300 pixels and an aspect ratio from 0.4 to 2.5. Each reference video is 3–30.05 seconds and up to 300 MB, with both sides from 700 to 4,553 pixels and 23.9–60 FPS; up to 5 videos, together within 30.05 seconds. An element uses 2–4 photos or one 3–8 second MP4 or MOV character video with a shortest side of at least 700 pixels, and that video counts toward the same 5-video, 30.05-second total. A voice is an MP3, WAV, M4A, MP4, or MOV file of 5–30 seconds, up to 15 MB. Names are 1–20 characters. Tag frames with @StartFrame, @Middle1, and @EndFrame, photos and videos with @Image1 and @Video1, and an element or standalone voice with its name, such as @Hero. A voice saved on an element is mentioned with that element. A reference video can keep its original audio. Prompts stop at 8,000 characters.

Each video is a whole number of seconds from 3 to 30, at 720p or 1080p. Native stereo stays on: there is no audio toggle, and audio does not add a separate credit line. Text uses 16:9, 9:16, 1:1, or 21:9. Keyframes follow the start photo. Reference to video can use auto, 16:9, 9:16, 1:1, or 21:9.

A 3-second 720p video is 26 credits, was 33. The default 5-second 720p video is 43 credits, was 55. A 30-second 1080p video with no reference video is 384 credits, was 494. A 10-second 720p video with 10 seconds of reference video is 103 credits, was 132. The top of the table is 461 credits, was 593, for 30 seconds at 1080p with 30 seconds of reference video. Photos do not change the quote. Reference-video seconds, including a video element, do. Kling 4.0 is on the Free, Creator, and Pro plans when the balance covers the quote. Generation is not zero-credit. Confirm the reservation before you generate. Failed runs return reserved credits. Cash refund requests are reviewed under the refund policy. Compare plans on the pricing page.