Kling 3.0 720p or 1080p: the same price per second on Sume

Most video models make you pay for pixels. Wan 3.0 is four times dearer at 1080p than at 480p, and Seedance 2.5 costs 2.46 times as much at 1080p as at 720p. Kling 3.0 on Sume is the exception: the price has no resolution term. If you use kling-3, there is no cost reason to pick 720p over 1080p.

The pricing code multiplies the number of seconds by one of two rates, depending on generate_audio. The rates are the provider list rates, $0.112 a second with audio off and $0.168 with audio on, times Sume's 1.25. The resolution field does not enter the calculation.

What a clip costs

Kling 3.0 on Sume takes 4 to 15 whole seconds at 720p or 1080p, in 16:9, 9:16 or 1:1. That makes the price table short.

The request

Send "resolution": "1080p" and the price stays on the same line. The poll response will report the same usage.cost you would have seen at 720p. A 720p request is still useful when your output is a phone-sized vertical that will be shown small, because it keeps the file smaller, but the saving is in bytes and not dollars. The price line is the same either way.

If you also need references, Kling 3.0 is the wrong id: it rejects reference_*_urls. It does support text-to-video and image-to-video with a start and an end frame.

How it compares at 1080p

Because Kling is flat, its position against the others changes with resolution. At 720p, Wan 3.0 is $0.125 a second and cheaper than Kling's $0.14 with audio off. At 1080p, Wan is $0.25 a second and Kling is cheaper on both settings. A 5-second clip at 1080p costs $0.70 on Kling with audio off, $1.25 on Wan 3.0 and $7.11 on Seedance 2.5 in 9:16.

The audio setting is the other lever, and it is switched on by default. Check what a clip without sound is worth to you before leaving it on, because the audio-on rate is 50% higher and applies to every second of the clip, whether or not anyone ever unmutes it.