What Claude actually costs a small business

Claude runs from $0 to $200 per person per month depending on plan and usage, and every paid tier includes Claude Cowork and Claude Code. The table below is the full price list as of August 10, 2026.

Two structural facts matter more than any single number. Team plans start at two seats, so there is no big minimum to clear. And standard and premium seats mix on one bill, so paying $100 seats for people who need $20 seats is a choice, not a requirement.

The Team plan on standard seats is the right default for a small business, and the individual plans are the wrong tool for one. Team adds the things an owner actually needs: one bill, admin controls, single sign-on, and the no-training default on your content.

Buy premium seats only for the one or two people who run Claude all day, and let usage tell you who they are. Starting everyone on standard and upgrading the seats that hit limits is cheaper than guessing up front.

The Max plans are personal plans, not team plans. They fit a solo owner or a single power user without a team account, and a team should get the same capacity through premium seats instead, on the shared bill with shared controls.

Offices weighing Claude against what they already pay for should read the seat price next to the alternatives: Claude vs Microsoft 365 Copilot for a small office covers the add-on arithmetic for a Microsoft shop, and Claude vs ChatGPT for business covers two products whose headline seat prices match to the dollar.

Between $2,400 and $4,320 a year covers the realistic range for ten people, depending on billing cycle and seat mix. The arithmetic, at the prices checked August 10, 2026:

For scale: a five-person office on annual standard seats is $100 a month, $1,200 a year. Against what a small business pays for most line-of-business software, the subscription is not the number that should worry anyone.

Usage headroom, mostly. A premium Team seat carries about five times the usage of a standard seat; Max 5x and Max 20x carry about five and twenty times Pro. The features are broadly the same tier to tier; what you are buying is how much Claude a person can use before hitting limits, plus priority during peak hours on Max.

The tell that someone needs a bigger seat is simple: they hit usage limits during normal work. Until that happens, nobody does.

Enterprise adds the controls that auditors and regulated clients ask about: audit logs, custom data retention, automatic account provisioning, network restrictions, and a HIPAA-ready configuration option, as of August 2026. Each control is unpacked in plain terms in Is Claude safe for client data? Pricing is a per-seat base plus usage that scales with model and task, quoted through sales.

It is worth it when someone outside the business can demand proof: a regulator, an insurer, a client audit, a bar association inquiry. A law firm handling sensitive matters often lands here; the reasoning is worked through in Claude for law firms. A ten-person office with none of those pressures usually does not need it on day one.

The real cost of Claude in a small business is the rollout, not the seats: choosing workflows, connecting tools safely, training people on real work, and writing the rules. Bought as staff time or bought as help, that work is the difference between seats that get used and seats that lapse.

• Setup: plan choice, accounts, access rules, connecting your document store or practice system.
• Training: sessions on your actual documents and workflows, not generic prompt tips.
• Governance: a written AI policy covering allowed data, review, and disclosure, signed before go-live.
• Measurement: a baseline before the rollout, so time saved is a number and not a feeling.

That list is our consulting engagement. One flat fee from $3,500 covers setup, training, and the signed policy for a team of 5 to 25, quoted in writing before any work begins. Details are on the pricing section of our main page.

Seat plans are flat, which does most of the work: a standard seat costs $20 a month billed annually whether it drafted three emails or three hundred. The variables left are seat count, seat mix, and billing cycle, and all three are decisions you revisit rather than surprises you absorb.

• Bill annually once you are sure, and pocket the roughly 20 percent difference.
• Review the seat mix after the first month of real usage, and upgrade only the seats that hit limits.
• Keep the API out of the mental model unless you are building software; it is metered per use and is a different product from staff seats.