Loading lab environment
Model access through a gateway with a per-learner spending cap.
2 vCPU, 4 GB RAM, API gateway, retrieval store and scoped tools. Per-learner API spending limit; mock mode when credits are unavailable.
This describes what the environment is designed to be. It has not been built, so these figures are a target rather than a measurement.
Runtime would be charged from the moment an environment is ready, never while it is queued and never for a failed start. These are design proposals and no billing exists.