Claude Code for teams costs $20 per standard seat per month on annual billing, or $100 per premium seat, and Anthropic lists Claude Code and Cowork against the premium seat. The Team plan runs from 2 to 150 people and you can mix the two seat types. Enterprise works differently: the seat fee buys access only, and every token your team uses in Claude, Claude Code and Cowork is billed separately at API rates with no included allowance and no cap. For a team of eight with two heavy users, mixed Team seats come to $3,840 a year. The same eight on Enterprise is $14,400 to $24,000 a year in usage alone, before the seat fee.
That gap is the whole decision, and almost nothing written about this plan spells it out. What follows is every published figure, read from Anthropic's own pricing page and help articles in September 2026, plus the arithmetic that the pricing page leaves to you.
What Anthropic actually charges a team
There are three shapes, not a ladder. Individual plans (Pro and Max) are consumer subscriptions with no admin controls. The Team plan is the seat-based product with two seat types. Enterprise switches billing models entirely.
On the Team plan, Anthropic publishes a standard seat at $20 per seat per month billed annually, or $25 billed monthly, described as "All Claude features, plus more usage than Pro". The premium seat is $100 per seat per month annually, or $125 monthly, described as "5x more usage than standard seats", and the pricing page lists "Includes Claude Code and Claude Cowork" against it. The plan is sold "For teams of 2 to 150", so the old five-seat floor people still quote is not what the page says today. Seat types can be mixed inside one organization, which matters more than it sounds.
Note what Anthropic does not publish. There is no numeric multiple for a standard seat against Pro, only the phrase "more usage than Pro". Every specific figure you will find elsewhere, the 1.25x and 6.25x numbers in particular, is a third-party estimate rather than something Anthropic prints. If you need a number in a spreadsheet before you commit, that absence is itself a finding.
Every 5x step Anthropic sells is priced at exactly 5x
Line the plans up and a pattern falls out that no vendor page draws attention to. A standard Team seat is $20 a month on annual billing. A premium seat carries five times that seat's usage and costs $100, which is exactly five times the money. On the individual side, Pro billed monthly is $20 and Max 5x carries five times Pro usage for $100, again exactly five times the money.
Both of Anthropic's five-times steps are priced at precisely five times. You are not buying a volume deal at any point; you are buying capacity at list. The one place the arithmetic breaks in the buyer's favour is Max 20x, which carries twenty times Pro usage for $200 a month, ten times the monthly Pro price. We ran that ladder out properly in the Claude Code Max plan pricing breakdown, where Max 5x works out at $240 per unit of Pro-equivalent usage against $120 for Max 20x.
For a team lead this has one practical consequence. Do not reach for premium seats because they look like the "team" answer. Reach for them when a specific developer is losing hours to a cap, and buy standard seats for everyone else. Mixing is supported and it is almost always cheaper than moving the whole organization up a tier.
Team against Enterprise: where billing flips from capped to metered
This is the part that catches finance teams out. Anthropic's Enterprise plan uses a single seat type, priced per user per month and billed annually, and its own documentation is unambiguous about what that fee covers: "The seat fee only covers access to the platform and doesn't include any usage. All usage across Claude, Claude Code, and Cowork is billed separately at standard API rates, based on what your team actually consumes. There are no per-seat usage limits and no included token allowance."
So Enterprise removes the cap, and removing the cap is exactly what makes it expensive. Anthropic publishes its own numbers for what Claude Code consumes at API rates: "around $13 per developer per active day and $150-250 per developer per month", with costs staying below $30 per active day for 90% of users. Multiply that across a team and the usage line dwarfs the seat line.
What you get for it is real, and it is compliance rather than capacity: audit logs, SCIM provisioning, custom data retention controls, a Compliance API that reaches activity logs, chat histories and file content with filtering by user and time range, an Analytics API, customer-managed encryption keys held in your own cloud, US-only inference, and a HIPAA-ready configuration with a BAA for eligible organizations. Admins can set spend limits at both organization and individual user level, which on an uncapped plan is not a nice-to-have. Enterprise can be bought self-serve or through sales.
The honest rule: buy Enterprise when an auditor or a regulator is the reason, not when your developers are hitting limits. If limits are the problem, premium seats fix them for a known number.
What this costs a real team of eight
Take an engineering team of eight where two people live in the agent all day and six use it in bursts. Annual billing throughout.
Mixed Team seats: two premium at $100 and six standard at $20 comes to $320 a month, or $3,840 a year. Capped, predictable, with admin controls and central billing.
All premium Team seats: eight at $100 is $800 a month, or $9,600 a year. You would only do this if all eight are genuinely heavy, and most teams discover they are not.
Eight individual Max 5x plans: also $800 a month, but Max is a consumer plan sold monthly only, with no admin controls, no central billing, no shared workspace and no way to reassign a seat when someone leaves. It is the same money for less governance, and it is the option teams drift into by accident when developers expense their own subscriptions.
Enterprise: the seat fee plus eight developers at $150 to $250 a month of metered usage, which is $1,200 to $2,000 a month in consumption alone, or $14,400 to $24,000 a year before you have paid for a single seat. Worth it for the compliance surface. Not worth it for headroom.
The spread between the cheapest sane option and the compliance option is roughly six times, and none of it is visible on a pricing page. If you want the same comparison across vendors rather than across Anthropic's own tiers, the real cost per developer of an AI coding agent puts subscription, credit and token billing on one page.
One thing that quietly wastes a team's allowance
On Pro, Max and Team seats the quota is shared across everything Anthropic sells. Claude Code in the terminal, Claude Code in VS Code and JetBrains, Claude on web, desktop and mobile, and Cowork all draw from the same pool. Anthropic states it plainly: usage limits are "shared across Claude and Claude Code, meaning all activity in both tools counts against the same usage limits".
In practice a meaningful share of a developer's Claude usage is not code at all. It is meeting notes, status updates, drafting an email to a stakeholder, summarizing a spec. That work is legitimate, but it is competing with the build for the same allowance, and on a $20 standard seat it is competing hard. Moving the scheduling and document side of the day onto an assistant built for office admin work leaves the coding allowance for coding, which is usually cheaper than upgrading the seat that ran out. If you want the mechanics of how the two windows drain, how the five-hour window and the weekly cap actually work covers it.
How much is Claude Code for teams?
$20 per standard seat per month on annual billing, or $25 monthly, and $100 per premium seat annually, or $125 monthly. Anthropic lists Claude Code and Cowork against the premium seat, and the plan covers 2 to 150 people. Enterprise is priced per user per month, billed annually, with all usage billed on top at API rates.
Does the Claude Team plan include Claude Code?
Anthropic's pricing page lists "Includes Claude Code and Claude Cowork" against the premium seat, and its Enterprise documentation still refers to legacy seat names of "Chat" and "Chat + Claude Code" for organizations that have not migrated. If you are budgeting for developers who need Claude Code, price the premium seat at $100 per seat per month on annual billing and treat anything cheaper as a bonus you should confirm at checkout rather than assume.
What is the minimum number of seats for the Claude Team plan?
Two. Anthropic's pricing page sells the Team plan "For teams of 2 to 150", so the five-seat minimum and the $125 monthly floor that older articles still quote are not what the page says in September 2026. There is a ceiling as well as a floor: above 150 people you are on Enterprise whether or not you wanted its billing model.
Should a small team buy Max plans instead?
Only if you are one or two people and governance genuinely does not matter yet. Max is a consumer subscription: monthly only, no admin console, no central billing, no seat reassignment, and no contractual no-training guarantee attached to an organization. At two people you are already inside the Team plan's range, and a mixed Team setup costs the same or less while giving you an org to manage. The moment a third person joins, the Max route stops making sense.
Where a repo agent fits instead
Every option on this page sells your developers more time at a keyboard, priced by how much of it they use. That is the right purchase when the work is exploratory and a human needs to be in the loop the whole way through. It is a strange purchase for the other half of an engineering backlog: the well-specified tickets nobody wants to sit through.
Agentcode is built for that half. You connect a GitHub or GitLab repository on any hosting, describe a task, and the agent plans it, edits the code, runs your existing test suite and opens a pull request for a human to read. It never merges on its own, so your review rules and your reviewers stay where they are. It costs $29 a month, flat, billed yearly: not per seat, not metered, and with no reset day to work around. For a team already paying $3,840 a year for capped seats, it is a different line item answering a different question. How the agent actually runs a task is the next page to read, and Claude Code pricing across every plan has the individual tiers in full.
All prices and quotations read from claude.com/pricing and Anthropic's own help and documentation pages in September 2026. Per-developer totals are our arithmetic on Anthropic's published figures, not prices Anthropic publishes.