I went back through openledger's ModelFactory documentation this week specifically to understand the royalty calculation because "earn OPEN every time your model is queried" is the headline, but the formula underneath it determines whether small specialized models can actually earn meaningfully or whether rewards concentrate at the top.
what's documented is this. every model deployed through ModelFactory becomes a Payable AI Model a smart contract that automatically distributes OPEN tokens to the developer based on usage metrics, relevance, and performance. no platform taking a cut. no approval process. the contract executes the payment directly.
what isn't documented is the weighting. usage metrics, relevance, and performance are three separate variables — but how they're weighted against each other isn't publicly detailed. does a high-query-volume general model earn more than a low-volume but high-performance specialized one? that formula determines everything about whether the long tail of niche model builders can compete. i went through the docs twice and couldn't find it. that's the part worth asking about.