A 30-second Kling 4.0 video without video input costs $4.275 at 720p or $6.375 at 1080p on PoYo's current list pricing. On fal's published text-to-video pricing, the corresponding estimates are $4.71 and $7.05. These are calculated generation costs, not a quote for a completed advertisement or a promise about usable output.
Prices were checked on October 24, 2026. PoYo values follow the current model pricing; fal values follow its text-to-video and reference-to-video pages. Account discounts, tax, currency conversion, and post-production are excluded.
Compare the same billing mode
| Input mode | Resolution | PoYo credits per billed second | PoYo USD per billed second | fal USD per billed second |
|---|---|---|---|---|
| No video input | 720p | 28.5 | $0.1425 | $0.157 |
| No video input | 1080p | 42.5 | $0.2125 | $0.235 |
| With video input | 720p | 17 | $0.085 | $0.094 |
| With video input | 1080p | 25.5 | $0.1275 | $0.141 |
For these PoYo tiers, one credit corresponds to $0.005. Images alone do not activate video-input pricing; a source video, reference video, or video-defined subject does.
PoYo's listed rates are approximately 9.2–9.6% lower than the matching fal rates in this table. This comparison concerns published rates and billing conditions. It does not establish a difference in model quality, availability, latency, or support.
Text and image requests: output seconds determine the estimate
Without video input, multiply requested output duration by the no-video rate:
| Output | PoYo 720p | fal 720p | PoYo 1080p | fal 1080p |
|---|---|---|---|---|
| 5 seconds | $0.7125 | $0.785 | $1.0625 | $1.175 |
| 10 seconds | $1.425 | $1.57 | $2.125 | $2.35 |
| 15 seconds | $2.1375 | $2.355 | $3.1875 | $3.525 |
| 30 seconds | $4.275 | $4.71 | $6.375 | $7.05 |
These calculations also cover image references and image-defined subjects when no video asset is supplied. Keep the resolution and requested duration identical when comparing platforms.
Video input: the lower rate applies to more seconds
PoYo calculates:
I = ceil(sum(input_video_durations))
O = output duration used for billing
billed seconds = O + max(I, O)
cost = billed seconds × with-video rate
Round only after summing the real input durations. fal describes input-plus-output billing with a minimum of twice the output duration; its page does not specify PoYo's exact fractional-duration rounding implementation. Use integer-duration examples for a clean cross-provider comparison.
| Video input → output | Billed seconds | PoYo 720p | fal 720p | PoYo 1080p | fal 1080p |
|---|---|---|---|---|---|
| 5s → 5s | 10 | $0.85 | $0.94 | $1.275 | $1.41 |
| 5s → 10s | 20 | $1.70 | $1.88 | $2.55 | $2.82 |
| 20s → 10s | 30 | $2.55 | $2.82 | $3.825 | $4.23 |
| 30s → 30s | 60 | $5.10 | $5.64 | $7.65 | $8.46 |
A five-second video reference does not make a ten-second generation cost ten times $0.085. The minimum is 20 billable seconds, so the estimate is $1.70 at 720p. Upload a video because it improves control over motion or identity, rather than to obtain a superficially lower per-second rate.
Fractional input and editing
Two PoYo reference clips of 3.2 and 4.1 seconds sum to 7.3 seconds, which rounds up to eight for billing. For a five-second output, billable duration is 5 + max(8, 5) = 13 seconds, or 221 credits / $1.105 at 720p. Real-duration validation still uses 7.3 seconds rather than eight.
For editing without an explicit duration, PoYo uses the ceiling of the source duration as the output billing duration while leaving the upstream request duration omitted. A 5.2-second source therefore gives six source billing seconds and six output billing seconds: 12 billed seconds, 204 credits / $1.02 at 720p. That calculation is not a promise that the returned file lasts exactly six seconds.
Budget for accepted footage, not the best attempt
Suppose a team makes ten 30-second 720p drafts without video inputs. The generation estimate is $42.75 on PoYo versus $47.10 on fal. If two drafts are accepted, the observed generation cost per accepted clip is $21.375 versus $23.55. The assumed 20% acceptance rate is an illustration, not a measured Kling success rate.
Track each attempt, its actual charge, review outcome, and later edits. A lower invoice for one attempt can be outweighed by additional retries. Start at the resolution and duration needed to check the hardest action, then spend more on a complete scene only after that works.
Frequently asked questions
Does a lower rate mean an edited video is cheaper? Not necessarily. Video input increases billable seconds. Compare complete requests using the formula above.
Are these creator subscriptions? No. These tables compare API list pricing. Creator-app subscriptions and promotions use separate terms.
What about failed or rejected tasks? Do not assume a universal refund or retry rule. Inspect the actual task charge and the provider's current terms before calculating production spend.
For request examples, read the API guide. For a production-sized exercise, use the 30-second product-ad workflow.