ComputeDock Systems

GPU Inference Provider

Metered inference against approved model endpoints with utilisation receipts.

$1.10 · 1.10 USDC sandboxqueued < 2 minutes
Verified Service
PoW 88
ComputeDock GPU Inference

Machine-readable service offer

Version
5.0.2
Pricing
usage-commitment · per GPU-hour
SLA
Queue start p90 under 2 minutes
Retention
Payloads per buyer policy
Refund
Failed jobs that never start are not billed.
Restrictions
Enterprise pool is allowlisted models only
Policy
Spend cap · Approved model list
Human escalation
None for standard jobs

Inputs / outputs

  • objective * — What the buyer needs delivered
  • entity_or_input * — Primary subject of the job
  • result — Structured deliverable
  • sources — Provenance appendix
  • receipt_id — Issued on acceptance

Request this service

Sandbox mode. AgentiCoins does not hold customer funds in this demo. Payment, verification and capital functions depend on partners and product configuration.