Kling O3 Standard Reference to Video: the stack is the brief (3s, 4s, 5s, 56cr)
Kling O3 Standard Reference to Video is reference-to-video. Upload the stills that must appear. Do not spend the whole budget because the slider goes that far.
Kling O3 Standard Reference to Video is reference-to-video. Upload the stills that must appear. Do not spend the whole budget because the slider goes that far.
Kling O3 Standard Reference to Video is Kling's reference-to-video row for character-consistent generation. Requires image: true. Audio: true. Durations: 3s through 15s. Credits: 56. Max output resolution: not listed. The stills are the brief. The agent upload for reusable characters and products is how those stills become a kit instead of a one-off drop.
References first, length second
Reference-to-video is not text-to-video with extra pictures in the sidebar. The faces, the SKU, the room you upload are what must appear. Character consistency is the job the description names. If the stack is three near-duplicates of the same three-quarter smile, you did not give the model a cast. You gave it a mood.
Requires image is true. No stack, no generate. Build the kit: one face, one product, one place. Then write the shot that uses them. A prompt that describes a fourth person the stack does not contain will invent, and invention is how Standard stops being consistent.
Aspects: 16:9, 9:16, 1:1. Pick one. Audio is true — name the line, the room, or the quiet.
56 credits is the first five seconds
The snapshot price is 56 credits. The price matrix is not a mystery: first 5s at 56, each extra second 12, up to 15s (max 168). The slider goes to 15s because the row lists 3s through 15s. That is not a suggestion to buy 15.
3s, 4s, 5s — the lengths in the title — are the honest Standard takes for a character beat. A 15-second hold of a face you already proved at 5s is how 56 becomes 168 for no new information. If the scene needs a second beat, cut. Two 5-second clips with the same kit beat one 15-second wander that burns extra seconds at 12 credits each.
No resolution token is on this row. Do not promise 4K in the brief because a different Kling line lists it.
The stack is cheaper than the extra seconds
Lock the kit once. Reuse it. That is the point of the agent flow above. Then spend 56 on a 5-second proof, not 168 on a first try.
The Kling roster has other O3 tiers and other jobs. This page is Standard reference-to-video only. Best reference-to-video model is the ranked cut if Standard is the wrong weight. Best Kling model is the family cut.
If you have no stills that must appear, you do not have a reference-to-video job. Pick text-to-video. If you have one still and you only want it to move, you have image-to-video, not a stack.
FAQ
What do I have to upload?
At least the images that must appear. Requires image is true. Category is reference-to-video. Character-consistent generation is the description. No stack, no row.
Why not always generate 15 seconds?
Because 56 credits covers the first 5s and each extra second is 12 more, up to 168. The slider is not the brief. Prove the cast at 3s–5s.
Does this row include audio?
Yes. Audio is true. Write the speech or the quiet. A silent prompt still gets a soundtrack you did not pick.
Is resolution listed as 4K or 1080p?
No max_output_resolution is set on this snapshot. Do not sell a resolution the row does not name. Sell the consistent character at 3–15s, 56 credits for the first five.