Claude · FinOps Profile

Claude Finops

FOCUS-aligned FinOps for Claude. Cost is split between per-seat consumer / business plans (Free, Pro, Max, Team, Enterprise) and pay-as-you-go API usage metered in input / output / cache tokens per model, plus server-tool charges (web search at $10 / 1k searches, code execution beyond the 1,550-hour monthly free tier at $0.05 / container-hour, Managed Agents session runtime at $0.08 / session-hour). Spend is gated by usage tiers (Tier 1-4 ceilings of $100 / $500 / $1,000 / $200,000 per month) until the organization moves to Monthly Invoicing. Effective unit cost is materially reduced by prompt caching (cache reads at 0.1x input price; cache reads excluded from ITPM rate limit on Claude 4.x) and by the Batch API (50% discount on input and output).

Claude Finops is the FinOps profile for Claude on the APIs.io network, aligned with the FinOps Foundation Framework.

It defines 9 billable meters, billed in USD, on a monthly cycle, and pricing category pay-as-you-go + subscription.

The profile maps 8 FOCUS columns for cost-allocation reporting.

Tagged areas include Artificial Intelligence, Generative AI, Large Language Models, FinOps, and FOCUS.

Category: AI Infrastructure Pricing: Pay-As-You-Go + Subscription Billing: Monthly FOCUS v1.3
Artificial IntelligenceGenerative AILarge Language ModelsFinOpsFOCUS

Framework Alignment

Framework
FinOps Foundation Framework
Data Spec
FOCUS v1.3

Charge Categories

UsagePurchaseTaxAdjustmentCredit

FOCUS Columns

BillingCurrency
USD
ChargeCategory
Usage
InvoiceIssuerName
Anthropic, PBC
PricingUnit
token
ProviderName
Anthropic
PublisherName
Anthropic, PBC
ServiceCategory
AI Infrastructure
ServiceName
Claude

Meters

tokens_input
Unit: token
Uncached input tokens billed at the model's base input rate (e.g. $5 / MTok for Opus 4.7, $3 / MTok for Sonnet 4.6, $1 / MTok for Haiku 4.5).
tokens_output
Unit: token
Output tokens billed at the model's output rate (e.g. $25 / MTok for Opus 4.7, $15 / MTok for Sonnet 4.6, $5 / MTok for Haiku 4.5).
tokens_cache_write_5m
Unit: token
5-minute cache writes billed at 1.25x base input.
tokens_cache_write_1h
Unit: token
1-hour cache writes billed at 2x base input.
tokens_cache_read
Unit: token
Cache read / hit tokens billed at 0.1x base input. On Claude 4.x models these do not count toward ITPM rate limits.
web_search_requests
Unit: search
Server-tool web searches at $10 per 1,000.
code_execution_hours
Unit: container-hour
Code-execution container hours beyond the 1,550-hour monthly free tier, at $0.05 / hour per container.
managed_agent_session_runtime
Unit: session-hour
Claude Managed Agents session runtime at $0.08 per session-hour, metered while the session is running (not idle).
consumer_seats
Unit: seat-month
Per-seat subscriptions for Pro, Max, Team, and Enterprise plans.

Sources