Build1 publisher2 min readPublished
Anthropic's usage-based Enterprise plan bills all tokens at API rates on top of the seat fee
The seat fee covers access and no tokens, so every prompt in Claude, Claude Code and Cowork bills to the organisation at API rates. The one control Anthropic's help page describes is an admin-set spend limit.
The Engineer · Build desk
What happened
- Anthropic's Enterprise help page says the seat fee covers platform access only, and that all usage across Claude, Claude Code and Cowork bills separately at standard API rates with no included token allowance.
- Enterprise uses a single seat type, priced per user per month and billed annually, giving each user Claude on web, desktop and mobile plus Claude Code and Cowork.
- Usage-based Enterprise plans have no plan-level or seat-level usage limits, and usage is billed on actual consumption instead of being capped at a fixed amount.
- Organisations on a seat-based plan with Standard and Premium seats do have usage limits, and can enable usage credits on those seats.
- Over the Team plan, Enterprise adds audit logs, SCIM, custom data retention controls, a Compliance API, an Analytics API, customer-managed encryption keys and US-only inference.
Compiled by The EngineerSomething wrong?How this is made
Why it matters
- decision The first question at renewal is which Enterprise shape you are on, because a capped seat-based plan and an uncapped usage-based one are forecast by different methods.
- constraint Cost control moves out of the contract and into the admin console, where a configured spend limit is what stops a month's charge from growing.
- capability Per-user activity is retrievable programmatically with time-range filters, so a budget owner can build internal chargeback instead of arguing about who ran the agents.
On usage-based Enterprise plans there is nothing to divide up. Anthropic's help page says one person's heavy usage does not reduce what is available to anyone else, because nothing is allocated in the first place [5]. In a pooled plan, one engineer running agents all week reduces what colleagues can do; on this plan the same week raises what the organisation owes [5].
Cost predictability, according to the same page, comes from spend limits that admins set at the organisation level or on an individual user [4]. The ceiling on a month's charge is whatever number an admin has set there [1].
The two APIs Enterprise adds are where spend attribution has to be built. The Compliance API returns activity logs, chat histories and file content, filterable by user and time range [9]. The Analytics API returns aggregated engagement and adoption metrics for the organisation [10]. Turning engagement counts into a per-team dollar figure means applying the API rate card yourself [4].
Connectors feed the same meter. Claude connects to Google Drive, Gmail, Google Calendar, GitHub, Microsoft 365 and Slack to search and retrieve context from existing documents, emails, calendars and team communications without manual uploads [17]. Retrieved context arrives as input tokens, and input tokens bill at API rates [3]. Switching on Drive and Slack retrieval changes the token count of every question that uses them.
Purchase mechanics matter here more than they usually do. Self-serve and sales-assisted Enterprise include the same features and the same seat pricing, and differ in how you buy, pay and manage the plan [12]. New self-serve organisations can choose ACH bank transfer, credit card or debit card at signup, while an organisation upgrading from a Team plan starts on credit card and can switch to ACH afterwards under Organization settings > Billing [13]. That leaves an uncapped metered charge running through a card until somebody changes it.
Invoicing, dedicated customer success management and payment in a currency other than USD all require a sales-assisted plan [15]. Sales-assisted invoices of $50,000 or more can only be paid by bank transfer, ACH or wire [14]. The help page does not list the per-seat price, and it says price and plans are subject to change at Anthropic's discretion [16][11].
A per-head budget survives this only if average consumption per user is steady enough to multiply by seat count [5]. Whether it does is a claim about your own workload, and you can only test it after the meter has run for a month.
What to watch
- Whether Anthropic ships a spend report or per-user cost breakdown next to the Analytics API, or leaves rate maths to the customer.
- Whether the seat-based plan with Standard and Premium seats stays available for buyers who need a fixed cap.
- Whether spend limits gain scopes beyond organisation and individual user, such as per team or per project.