The cleanest way to see the split is the two purest examples in the catalog, both self-evident from their own names. Per-second billing charges more the longer the clip runs, the way a taxi meter does. Flat billing charges the same number of credits for the job regardless of length, the way a fixed quote does. Between those two poles sits a third shape entirely: billing keyed to the size of the input or output rather than to time at all — a rate per megapixel of image, or per thousand characters of script read aloud.
The confusion this causes is specific and avoidable: people carry an assumption from the last model they used onto the next one without checking. Someone used to a per-second model assumes a longer duration option always costs proportionally more, and is surprised when a flat-billed model charges the same either way. Someone used to a flat-billed image model assumes size does not matter, and is surprised when a per-megapixel model charges more for a larger canvas.
None of this is a setting you choose on a given model — the billing type is fixed per model, decided by how the provider underneath actually charges, not by anything you select in the interface. Checking which basis a model uses before assuming how a longer duration or a bigger canvas will move the credit cost is the only reliable habit, because the catalog genuinely does not use just one.
In practice
- Check which basis a model uses before assuming a longer duration costs proportionally more — flat-billed models charge the same either way.
- Per-megapixel and per-character billing scale with output size or input length, not with a duration slider at all.
- Billing type is fixed per model; you cannot select a different basis on the same model, only pick a different model.
How the catalog meters a generation
Every model with a recorded billing rule, grouped by what the charge is actually keyed to. 303 of the 305 models in the Versely catalog qualify.
| Model | Provider | Type |
|---|---|---|
| GPT Image 2 Text to Image | OpenAI | Image |
| Mai Image 2.5 Edit | Microsoft | Image |
| Happy Horse 1.0 Text to Video | Alibaba | Video |
| Nano Banana 2 | Image | |
| Seedance 2.0 | ByteDance | Video |
| GPT Image 1.5 | OpenAI | Image |
| Gemini 3.1 Flash TTS | Audio | |
| Wan 2.7 Text to Video | Wan | Video |
Browse all 177 spec pages for full settings, resolutions and credit costs.
The mistake to avoid
Assuming every video model is priced like the last one you used. The catalog mixes over a dozen distinct metering rules side by side, and carrying an assumption from a per-second model onto a flat one, or a resolution-scaled one, misprices the job in your head before you have even generated anything.
Related terms
Credit
A credit is Versely's own metering unit — a fixed amount is deducted from your balance for each generation, so one number covers every provider's wildly different native pricing.
Duration
Duration is how long a generated clip runs, chosen before generation from whatever lengths the model supports rather than trimmed afterwards.
Resolution
Resolution is how many pixels an output contains, usually named by its height — 720p, 1080p, 4K — and set before generation rather than after.
Distillation
Distillation trains a smaller or faster model to imitate a larger one's outputs, which is where the fast and turbo variants of familiar models come from.
The all-in-one AI studio for creators. 60+ models for video, image, voice, music and lipsync in a single app.