How much does the Claude Team plan cost?

The Claude Team plan has two seat types: a Standard seat at about $25/month (roughly $20/month billed annually) and a Premium seat at $125/month ($100/month billed annually), the difference being usage headroom. Claude Code is included on every seat. Anthropic recently reduced the old five-seat minimum, so many accounts can now start with fewer seats — verify the current minimum at checkout.

Is the Claude Team plan worth it?

Team is worth it when a group needs shared projects, common connectors, centralized billing, admin controls, and SSO — plus Claude Code on every seat. If it's really just one heavy user, a $200 Max plan gives one person more usage with no per-seat minimum. The tiebreaker is collaboration and control, not raw usage: buy Team for the shared workspace, buy Max for individual capacity.

Is there a separate cost for the Claude API and models on top of a plan?

Yes. Beyond the subscription, using the API — or exceeding your plan's limits and enabling usage credits — is billed per token, and the rate depends heavily on the model, from Haiku 4.5 at $1/$5 per million tokens up to Fable 5 at $10/$50. As of July 7, 2026, Fable is billed at that metered rate on paid plans once its included allowance is used. Choosing and routing models to control that bill is its own topic, covered in our cost-first guide to which Claude model to use.

Does the Claude Team plan include Claude Code?

Yes. Claude Code is included with every Team plan seat, which is part of what makes Team a strong value for a small team that mixes developers and operators. The model usage that Claude Code consumes still counts against your plan's limits, so heavy automated use can push you toward usage credits — where the model-routing discipline (defaulting to Sonnet, escalating deliberately) keeps costs predictable.

July 7, 2026 · 12 min read

Claude Plans in 2026: Pro vs Max vs Team vs Enterprise — the Real Math

Q: What is the difference between the Team and Enterprise plans?

Enterprise is not simply Team with more usage — it adds the controls large organizations require: SAML SSO and SCIM provisioning, audit logs, a compliance API, custom data retention, role-based access, per-seat spend caps, HIPAA/BAA coverage, data residency, and custom contract terms. You move to Enterprise when you cross roughly 150 seats or when identity, security, or compliance requirements become non-negotiable.

Claude has two prices — the subscription and the model bill. The real math on Pro, Max, Team, and Enterprise: what each costs, when each one wins, and the seat details Anthropic keeps quietly changing.

ClaudeClaude pricingClaude Team planClaude EnterpriseAI tools

Claude has two prices, and most people only think about one. This guide covers the first: the subscription plans. Pro is $20/month, Max is $100 or $200/month, Team runs from about $25 to $125 per seat per month with Claude Code included on every seat, and Enterprise is custom. Which one to buy is a decision about usage headroom and team controls, not access to smarter models — every paid tier reaches the same model lineup. The second price is the model bill (what you pay per token once you use the API or pass your plan's limits), and it just became newsworthy: as of today, July 7, 2026, the subsidized Fable access bundled with paid plans ends. We cover that in a companion guide; here we do the real math on the plans — Pro vs Max vs Team vs Enterprise, when each one actually wins, and the seat details Anthropic keeps quietly changing.

We pay for most of these plans, run the API, and operate autonomous systems on Claude every day. So this is the pricing breakdown we wish existed: not the feature grid, the decision math.

The subscription plans, in one table

Anthropic sells four subscription tiers, and the jump between them is about usage headroom and team controls, not access to smarter models. Every paid tier can reach the same model lineup; what you're buying up the ladder is more capacity, collaboration features, and administrative control.

Plan	Price	Who it's for	Key point
Pro	$20/mo	One person, steady use	Baseline paid usage (roughly 1× the reference limit)
Max	$100/mo (5×) or $200/mo (20×)	One heavy individual user	The most raw usage per person; no team features
Team	~$25–$125 per seat/mo	A department or small company	Shared projects, admin, central billing, SSO; Claude Code included
Enterprise	Custom	Larger orgs, IT/compliance needs	SAML/SCIM, audit logs, HIPAA/BAA, data residency, custom terms

Two things on the Team row, because it's the one people get wrong. First, Team has two seat types: a Standard seat (around $25/month monthly, ~$20 annual) and a Premium seat ($125/month monthly, $100 annual). The difference is usage headroom — a Premium seat carries several times a Standard seat's allowance — and you can mix them across a team. Second, Anthropic has been actively changing Team: the old five-seat minimum was recently cut (many U.S. accounts can now start at two seats), and seat caps have crept up. Because these terms move, treat every number here as "verified on the publish date, confirm at checkout."

Team vs stacking Max accounts vs Enterprise

The honest decision: buy Max when one person needs maximum usage, buy Team when a group needs shared workspace and admin, and move to Enterprise when compliance and identity controls — not usage — become the constraint. Most of the confusion online comes from comparing these on price-per-seat, which is the wrong axis. Compare them on what problem they solve.

When Max beats Team. If it's really just you (or a couple of independent power users), a $200 Max 20× plan gives one person more usage than a single Team seat does, with no per-seat minimum and no admin overhead. Two people who don't need to share projects can run two Max plans and be done. Team earns its keep when you need the shared parts.

When Team beats stacking Max accounts. The moment you want shared projects, a common set of connectors, centralized billing, domain-based onboarding, admin visibility, and SSO through Google or Microsoft, Team is the right container — and Claude Code comes included on every seat, which quietly makes it a strong deal for a small dev-plus-operator shop. This is the exact comparison a widely-shared Reddit thread wrestles with ("Team plan vs just getting multiple 5× accounts"), and the tiebreaker is almost always collaboration and control, not raw tokens.

When Enterprise is worth it. Enterprise is not "Team with more usage" — it's Team with the controls IT and legal require: SAML SSO and SCIM provisioning, audit logs, a compliance/audit API, custom data retention, role-based access, per-seat spend caps, HIPAA/BAA coverage, data-residency options, and a real negotiated agreement. You reach for it when you cross roughly 150 seats, or when any one of those requirements is non-negotiable. If nobody in your org is asking about SCIM or a BAA, you're not ready to pay for Enterprise, and that's fine.

The other price: what the models cost (the short version)

The second price is the model bill — what you pay per token on the API, or once you pass a plan's limits and switch on usage credits — and the spread is enormous: Fable 5 costs ten times what Haiku does. The quick ladder: Haiku 4.5 ($1/$5 per million tokens) for high-volume simple work, Sonnet 5 ($3/$15, intro $2/$10 through August) as the everyday default, Opus 4.8 ($5/$25) for genuinely hard reasoning, and Fable 5 ($10/$50) for frontier, long-context work. Most work belongs on Sonnet; reaching for the top model by reflex is the fastest way to a surprising bill.

This matters more today than yesterday. As of July 7, 2026, the subsidized Fable 5 access bundled into paid plans ended, so continued Fable use is now billed as metered credits at that $10/$50 rate. If you use Claude in the app, the model choice is mostly handled for you and this barely touches you. If you build on the API or run automations, routing work to the right model — plus prompt caching and the Batch API — is the difference between a predictable line item and a scary one.

That decision deserves its own guide, so we wrote one: which Claude model to use — the cost-first guide covers the per-token math, the three levers that cut the bill, the honest take on who actually needs Fable (spoiler: almost nobody), and the routing map we run our own systems on. Those systems cost about $200/month partly because they never touch Fable — the same discipline behind the honest agency economics we've written about.

So what should you actually buy?

Most individuals and teams should live on a subscription and treat the API and Fable as opt-in tools, not defaults. If you're one person, Pro or a Max plan covers you and the model choice is handled inside the product. If you're a team that needs to share work, Team is the container, and Claude Code riding along on every seat makes it a better deal than it looks. Enterprise is a control-and-compliance decision, not a usage one.

The API and usage credits are where you go when you're building something — automations, products, loops — or when your usage is spiky enough that metered beats a bigger subscription. That's exactly where the model-routing discipline pays off, and exactly where today's Fable change will separate the teams who set a cap and route deliberately from the ones who get a surprising bill. If you want help figuring out which side of that line your operation is on, that's what our Revenue Audit is for, and if you're weighing Cowork specifically, our Claude Cowork pricing breakdown goes one level deeper on that product.

FAQ

Published: July 2026. All plan prices, model rates, and the Fable 5 usage-credit change (effective July 7, 2026) were verified against Anthropic's pricing pages and documentation on the publish date. Claude pricing changes frequently — confirm current numbers before you buy or enable usage credits.

Keep reading

← All insights Start a conversation →