Three rails. Bring one or both.
You can list a model, list compute, or bring both. Cautus Intelligence gates the call, meters tokens, and settles once on Litheum. We are not a GPU landlord and we do not take a second cut when pool compute is matched.
Three rails
Bring a model, compute, or both.
Stock, fine-tune, or specialist. Passport shows the class. You keep the weights.
Bring GPUs, join the pool, or let the buyer use theirs. Listing compute is free.
Open API, or approve / deny with competitor blocks. Access is not a fee.
Model
Classify the listing as stock, fine-tune, or specialist. Class is shown on the passport and only moves the creator royalty on matched jobs. It does not change the platform take.
Compute
Self-hosted GPUs, the shared pool, or the buyer’s own machines. Listing compute is free. Matched jobs pay the model owner a class royalty and the rest to the GPU provider.
Access
Open API, or request / approve / deny with optional competitor domain blocks. Access is a control. It is never a fee surcharge.
Who can show up as whom
- →Model only: list weights, match pool or third-party GPUs, 6% rail take.
- →Compute only: put GPUs in the pool. No listing take. Earn the compute share.
- →Both: same host brought model and GPUs, 5% rail take, 100% of remainder back.
- →Buyer BYO: 2.5% rail take; model owner gets the rest. Buyer already paid their GPUs.