As of 18 October 2026, Anthropic's pricing page lists the Claude Team plan for teams of 2 to 150, at 20 dollars per standard seat per month billed annually, or 25 dollars billed monthly. It publishes no absolute token allowance for seats, only relative usage. Prices and limits change, so check the page before you buy.
Everything here comes from the pricing page fetched on 18 October 2026. We make no claim about what the plan costs Anthropic or how limits are enforced internally.
What does a seat cost?
The page lists two seat types.
- Standard seat: 20 dollars per seat per month if billed annually, 25 dollars if billed monthly.
- Premium seat: 100 dollars per seat per month if billed annually, 125 dollars if billed monthly.
You can mix standard and premium seats on one team, according to the page. The stated team size range is 2 to 150.
What are the usage limits?
The page describes standard seats as having all Claude features plus more usage than the Pro plan. It does not give a number. Premium seats are described as giving 5 times more usage than standard seats. Again, the page gives a ratio, not an absolute figure.
The page also says usage limits apply, and its FAQ says Team limits may include weekly and monthly caps at Anthropic's discretion. No token counts for Team seat allowances are published there.
The page does list per million token rates for models under API pricing. Those are separate from what a Team seat includes, so you cannot convert a seat price into a token total from them.
What else is included?
The page lists, among other items, access to Claude Code and Cowork, Microsoft 365 and @Claude integrations, enterprise search, central billing, single sign on, connector admin controls and desktop deployment. It also says models are not trained on your content by default, and mentions up to 500 dollars per month in pooled API credits. Some items are labeled beta. Read the page itself for the current list, since it changes.
How do you plan around unpublished limits?
Since no token number exists to plan against, treat the first month as a measurement period.
- Start with a pilot. Put a few heavy users on a trial and watch how often they hit a limit.
- Segment by workload. People who run long coding sessions behave differently from people who draft email. The premium seat exists for the first group.
- Track blocked time. The cost that matters is time lost while someone waits for a limit to reset, not only the seat price.
- Compare with API pricing for automated work. For batch jobs and pipelines, per token rates are published, so costs there are predictable.
- Re check quarterly. Because the page can change, record the date of every price you rely on.
How does this relate to security tooling?
If your team uses an AI coding assistant, whatever plan it runs on, the code it writes needs the same checks as any other. Safeguard's MCP server at mcp.safeguard.sh can be connected to Claude, and tool calls to it are unlimited on every plan, so the security checks do not compete with your assistant's usage allowance on Safeguard's side. Do not read that as a statement about Anthropic's limits. Those are set by Anthropic.
What to do
- Open the pricing page and note the date and the figures you rely on.
- Decide how many premium seats you need by looking at who does long sessions.
- Run a two to four week pilot and log limit events.
- Budget for the seat mix you measured, not the one you hoped for.
- Review again when the page changes.
Quick check
How many tokens does a Team seat include?
The page we fetched on 18 October 2026 does not say. It gives relative statements only, such as premium seats having 5 times the usage of standard seats.
Is the monthly price the same as the annual price?
No. The page lists a lower per seat price when billed annually and a higher one when billed monthly.
Sources
- Anthropic: Claude pricing, fetched 18 October 2026
Last reviewed: 18 October 2026.
