Providers do not price alike. One bills by the second of finished video, another by a flat fee regardless of length, another by the megapixel of an image or the thousand characters of a script. A customer trying to compare a job on one model against a job on another would otherwise be comparing seconds to megapixels to characters — three units with nothing in common. A credit exists to collapse all of that into one number: whatever the provider actually charges by, the job costs a set number of credits, decided up front.
A credit is routinely confused with a token, and the two measure different things entirely. A token is a unit of text — the currency an LLM's context and output are counted in, sized to roughly a word fragment. A credit is a unit of a whole job: one generation, however long it took the model to think or how many characters it read, priced as a single number decided before the job runs rather than metered by anything counted during it.
On Versely's own catalogue that floor is 1 credit — a handful of fast image generators and utility passes charge exactly that — and the recorded ceiling for a single generation currently sits at 210, on the most demanding video jobs. Per-option pricing on models that scale by resolution and duration together can run higher again once every size and length combination is counted, which is exactly the billing-type question the next entry answers.
In practice
- Credits are deducted per job, not per token, character or second directly — that arithmetic already happened before the price was set.
- The same brief costs a different number of credits on different models; a credit buys a finished job, not a fixed slice of compute.
- Batch generation multiplies linearly — ten variants of one brief cost ten times what one costs.
The mistake to avoid
Comparing credit costs across models as if a credit means the same thing everywhere. A credit is Versely's price, not a fixed amount of compute — two models at an identical credit cost can differ hugely in resolution, duration or what the provider underneath is actually doing; compare the finished job, not the sticker number.
Go deeper
Credits Explained: Budgeting AI Content Like a Media Buyer
How Versely credits work and how to budget them like a media buyer: cost drivers, model tiering, draft-vs-final strategy, and a monthly planning framework.
Related terms
Billing type
Billing type is which measurable thing a model's price is actually keyed to — time, a flat per-job charge, or the size of the input or output — and it is why two models can charge wildly different credit totals for what looks like the same request.
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.
Batch generation
Batch generation runs one brief as several jobs at once — multiple variants, models or formats — so the comparison happens in a single sitting instead of across a week.
The all-in-one AI studio for creators. 60+ models for video, image, voice, music and lipsync in a single app.