BVINF — Inference, tokenized per call
Buy inference capacity in bulk. Resell unused. Hedge per-call cost. Each BVINF is one verified inference call against a registered BVMDL model, with TEE-attested execution.
The trade
If BVMDL is the IP rail, BVINF is the run-time rail. Each BVINF is one paid inference call, attested by the serving runtime, redeemable against the registered model.
Mechanics
- 1 BVINF = 1 inference call against a registered BVMDL
- TEE attestation of execution (compute provider's signed proof)
- Settle: MXD on Arbitrum
- Resale: BVINF is a transferable ERC-20 — unused capacity tradable
- Pricing: per-model curve set by the BVMDL creator
More from BV Compute
MXTK API Gateway
TestnetApollo Federation GraphQL gateway + REST surface. Single query surface across mortgages, tranches, energy tokens, tax credits, compute markets.
BVGPU
TestnetGPU-hour token. Each BVGPU represents one verified hour of qualified GPU compute (H100 / A100 / equivalent), redeemable for delivery or cash settlement.
BVMDL
TestnetModel token. Each BVMDL represents licensed access to a specific AI model weight set or fine-tune, with usage metering and revenue share.