What the fleet actually costs.

Every deliberation, sign-off, and triage the iris fleet runs is metered against real xAI rates. Deliberation USD is exact (prompt/completion split × per-model rate); sign-off USD is estimated at the input rate (prompt-heavy). Build cost is metered per pipeline run. No fabricated pricing.

Loading… · auto-refresh 5 min

Ownership economics · rented vs owned

External vs internal — and what we keep.

Every metered surface, split three ways: dollars that actually left CH for rented inference, the rented value of that same work, and the dollars kept by running on CH-owned compute. Savings begin accruing the day local models come online — watch this line cross from rented toward owned.

Rented vs owned — cumulative, over time

Loading…

Deliberation spend by model

Model	Runs	Tokens	USD	Rate
Loading…

Source: deliberation_costs + fleet_signoff + pipeline_run (build cost, tokens exact) × ai_model_pricing (real xAI rates: grok-4.x $1.25/$2.50 base, >200K $2.50/$5.00; grok-build-0.1 $1.00/$2.00). Advisory-consult tokens not yet metered.
Integrity In All We Do℠ · Conceptual Healthcare Corporation