Claude Team vs Enterprise Pricing: Seat Cost, Minimum Seats and Claude Max vs Team

Claude Team costs $20 per Standard seat or $100 per Premium seat per month on annual billing and supports 2 to 150 seats, with usage included in the seat. Claude Enterprise costs $20 per seat billed annually for access only, with every token billed separately at API rates, and starts at 20 seats self-serve or 50 seats through sales. The seat price is identical on Standard Team and Enterprise. What changes is whether there is a ceiling on the bill.

That is the whole decision in three sentences, and it is worth reading twice, because the identical $20 is exactly what trips people up. A finance team that migrates 60 people from Team to Enterprise on a like-for-like seat count sees no change in the line they were watching and then meets a second line they were not.

Claude Team vs Enterprise pricing, side by side

Everything in this table is published by Anthropic on claude.com/pricing or support.claude.com, and was checked on September 6, 2026.

Which plan can you actually buy?

Before comparing anything, check which plans your headcount even permits. Three published thresholds carve the market into bands, and two of them are easy to miss.

The 20 seat self-serve Enterprise floor is the surprising one. You do not need a procurement process or a sales call to buy Claude Enterprise. At 20 seats or more you can sign up online at claude.ai/create/enterprise with a card or ACH transfer in US dollars, and you get audit logs, SCIM provisioning, custom data retention controls, customer-managed encryption keys, compliance and analytics APIs, US-only inference options and HIPAA eligibility. A 25 person company can have all of that running the same afternoon. The 50 seat minimum applies only to the sales-assisted path, where you are buying custom terms rather than the standard product.

The same $20 buys two completely different things

A Team Standard seat is a bundle. The $20 includes a usage allowance, measured as 1.25x the session usage of an individual Pro plan, on a weekly limit that applies across all models and resets on a fixed day assigned to your account. When one member exhausts it, everyone else keeps working, because limits are per seat rather than pooled.

An Enterprise seat is a key. Anthropic's own help center says the seat fee "covers access only, and all usage is billed separately at API rates." Those rates are the public ones: $5 and $25 per million input and output tokens on Opus 5, $2 and $10 on Sonnet 5, $1 and $5 on Haiku 4.5.

Put a number on what that second line means. At a 90% input to 10% output token mix, $20 of Sonnet 5 usage is about 7.1 million tokens, and $20 of Opus 5 usage is about 2.9 million. Anthropic's own support-agent worked example prices 10,000 conversations at roughly $37.00 on Haiku 4.5, so one $20 seat fee is worth about 5,400 automated conversations in tokens. For light chat use the access fee dominates. For anything agentic, it disappears in the first week and the usage line is the only number that matters.

Anthropic sells no volume discount at all, and the arithmetic proves it

Here is something neither plan page states and every buyer assumes the opposite of. Anthropic describes seat capacity as a multiple of Pro rather than in tokens, which means you can divide price by capacity and compare the plans on a single unit. Do that across the whole monthly ladder:

Every monthly plan Anthropic publishes lands on exactly $20.00 per unit of usage. Not approximately. Exactly. Buying six times more capacity costs six times more, and buying it in one seat instead of five saves nothing.

One caveat on the Max row: Anthropic publishes "From $100" per month for Max and offers a choice of 5x or 20x Pro usage, so $100 is the entry price against the 5x tier. The 20x price is not published, and it may well break the pattern. Everything else in that table is a stated price against a stated multiple.

Switch to annual billing and the constant moves but stays constant. A Team Standard seat is $20 for 1.25x, which is $16.00 per unit. A Team Premium seat is $100 for 6.25x, which is also $16.00 per unit. Pro is $17 for 1x, so $17.00. Annual commitment is the only discount Anthropic sells, and it is worth 20% on Team seats against 15% on Pro.

Two things follow that will change how you buy. Choosing between Standard and Premium seats is not a cost optimization, it is a capacity decision, so give Premium to whoever keeps hitting the weekly wall and stop treating it as a tier to negotiate. And a solo operator paying $100 a month for Max is paying $20.00 per unit while a colleague on an annual Team Premium seat pays $16.00 for more headroom. Team needs two seats minimum, so one person cannot buy their way into that rate alone, but two people can.

The spend ceiling is the difference nobody quotes

Both plans sell extra usage once a seat's allowance runs out, and both price those credits at standard API rates. The direction of the billing is where they split.

On Team, owners "pre-purchase usage credits that they can control using spend limits." You buy a block, set a cap, and the product enforces it. Your worst case is known before the month begins.

On seat-based Enterprise, usage credits "are billed at the end of each month based on your actual usage during the billing period." You consume, then settle. There is no ceiling in that sentence, which means the ceiling has to be yours.

For most finance teams this is the whole argument, and it has nothing to do with the seat price. If you are moving off Team, the action item is to replace the cap you are giving up before the migration rather than after the first invoice that makes someone's stomach drop. Folding the Anthropic line into whatever you already use to watch SaaS and cloud spend read-only gets the alerting back, since the plan change removes the in-product version of it.

When is it worth upgrading from Team to Enterprise?

Four triggers justify the move, and only one of them is about price.

You are approaching 150 seats. This is not a judgment call. Team stops at 150 and there is no negotiating past it, so plan the migration a quarter before you hit the wall rather than during a hiring spike.

You need identity and audit controls. SCIM provisioning, audit logs, custom data retention and customer-managed encryption keys are the reason most 60 to 120 seat organizations move early. If your security review has a line for automated deprovisioning, Team will not pass it.

You have a data residency requirement. Enterprise carries US-only inference options. Price it before you sign: setting inference to US-only applies a 1.1x multiplier to every token pricing category on Claude 4.6 and later, so it adds a flat 10% to your entire usage bill, not a fixed fee. A team burning $10,000 a month in tokens pays $12,000 a year for the residency guarantee alone.

Your usage is genuinely heavy and mostly agentic. This is the only case where Enterprise can be cheaper rather than merely more capable. Unmetered access at API rates, combined with prompt caching at 0.1x input and the Batch API at 50% off, can take Sonnet 5 input as low as $0.10 per million tokens on batch cached work. A well-engineered pipeline will beat a seat allowance. A naive one will not come close.

If none of those four is true, stay on Team. You are giving up a spend cap in exchange for features you are not using.

What Anthropic's own pages disagree about

Worth knowing before a negotiation. The pricing page and the help center describe Enterprise differently, and both were live on September 6, 2026.

claude.com/pricing shows Enterprise as "$20/seat" with "Seat price + usage at API rates," and the Enterprise help article says the seat covers access only. That is a plan with no included usage. But a separate help article about managing usage credits, titled for "Team and seat-based Enterprise plans," says credits let members keep working "after reaching their included usage limits." Included limits, on Enterprise.

The likeliest explanation is that Anthropic sells more than one Enterprise shape and does not name them apart on the pricing page. Do not resolve it by guessing. Get three lines into your quote: whether your seats include any usage, whether credits are pre-purchased or settled in arrears, and which rate card your usage prices against, because the published API list changes independently of your contract.

The short answer

Under 20 seats, the question is moot and Team is your plan. Between 20 and 150, stay on Team unless you need SCIM, audit logs, data residency or HIPAA, because the spend cap you keep is worth more than the seat price you do not save. Above 150, Enterprise is the only door and the job is to rebuild your ceiling outside the product before you walk through it.

If ChatGPT is on the same shortlist, the decision runs almost identically on OpenAI's side, down to the same four seat prices, and we cover it in ChatGPT Business vs Enterprise pricing.

And if what pushed you into this comparison was one repeatable job rather than a rollout, price that job on its own. A seat license is the right unit for people who will use an assistant differently every day, and the wrong unit for invoice reading or ticket triage where nobody is sitting there. Our full breakdown of Claude Enterprise pricing has the seat and API math in one place, AI agent platforms compared puts eleven vendors' rates side by side, and how AI agent pricing works covers the three metering models you will meet on any shortlist.