Model pricing reference
Revix bills chat turns from the actual cost of the work performed. Input, output, thinking output, and cached input can each have a different rate. A billed turn costs at least 0.1 credit.
| Model | Plan | Thinking | Relative price |
|---|---|---|---|
| Revix | Every plan | Off for simple edits; low otherwise | 1/3 |
| Revix MAX | Developer and Studio | Off for simple edits; low otherwise | 2/3 |
One credit covers $0.0125 of metered AI cost. Provider cache discounts are included; if the provider omits its cost, the fallback token rate follows the selected model's estimated price.
The model picker displays a current starting estimate when available. Actual credits depend on the length of your request and response, how much thinking the task needs, and cached context. Rates can change when Revix switches between its reviewed service providers.
See Choosing a Model and How Credits Work.