Interest Rate Models

Every VToken stores its own interestRateModel() address. Markets on the same network or in the same pool can use different model families and parameters, so never select a model from the network name alone.

The v4.4.0 engine source contains three implementations:

Model	Curve	Parameters mutable after deployment?
JumpRateModelV2	one kink	yes, through AccessControlManager
TwoKinksInterestRateModel	two kinks, signed slopes	no; constructor immutables
WhitePaperInterestRateModel	one linear slope	no; constructor immutables

BaseJumpRateModelV2 is not a separate contract in the current source. The old page is retained only to route existing links to JumpRateModelV2.

Units and annualization

The shared interface returns a mantissa-scaled rate per slot. A slot is either a block or one second, according to the deployed model's isTimeBased() value. Read blocksOrSecondsPerYear() from that same model before annualizing.

nominal annual rate = rate per slot × blocksOrSecondsPerYear

Do not assume 2,628,000 blocks/year or 31,536,000 seconds/year globally. In particular, the live BNB Chain and opBNB VToken implementations use different block constants; the engine version map records their current values.

Selecting the model for a market

1. Start from the market's VToken address in the deployed markets registry.

2. Read interestRateModel() from the VToken at the block used by your calculation.

3. Identify the model bytecode or verified source, then read its clock mode and parameters.

4. Preserve the effective block when recording a mutable JumpRateModelV2 parameter set.

The formulas and current ABI below follow the isolated-pools v4.4.0 source.