What is credit-based pricing?
Paying per action instead of a flat fee — and why credits tend to disappear faster than the number on the page suggests.
Credit-based pricing charges for usage in units — credits — spent per action: a generation, an enrichment lookup, an export. You get or buy a set number each period, and different actions can consume different amounts, which is exactly why total monthly cost is harder to predict upfront than a flat subscription fee.
A worked example
1,000 monthly credits, spent on a real task
| Action | Cost | Quantity | Credits used |
|---|---|---|---|
| Standard image generation | 5 credits | 40 images | 200 |
| High-resolution upscale | 15 credits | 20 images | 300 |
| Batch variations | 10 credits | 50 variations | 500 |
| Total | 1,000 / 1,000 |
The 1,000-credit plan sounds generous until the higher-cost actions — upscaling, batches — eat through it far faster than simple generations alone would. Reading the per-action cost table matters more than the headline credit number.
Why tools price this way
Different actions genuinely cost the provider different amounts of computing power — a flat subscription would either overcharge light users or undercharge heavy ones. Credits let cost track actual usage intensity, at the price of making your own monthly bill harder to predict in advance.
Related terms
FAQ
Do unused credits roll over to the next month?
It varies by provider — some reset unused credits monthly, others allow rollover up to a cap. Always worth checking the specific tool's policy.
Why do some actions cost more credits than others?
Because they require more computing resources — higher resolution, longer processing, or more complex generation typically costs more.
Can I buy extra credits without upgrading my plan?
Many tools offer this as a top-up option, separate from changing your base subscription tier.
Back to the full AI Glossary, or compare this against flat pricing in What is Freemium?