How much does Claude Code cost for a team?

The plans, side by side

Anthropic's list prices in US dollars before tax, from claude.com/pricing in October 2026. On Pro, Max and Team, usage comes out of limits that reset every five hours and every week, so the bill stays the same and a heavy week ends in a wait instead of a bigger invoice, unless you turn on usage credits. On the API and Enterprise's usage billing there are no such limits, only rate limits, and the bill moves with use.

What the API costs per token

Claude Code sends the whole conversation with every request, and each tool call sends another, so most of the tokens a session uses are its own history, read again. The prompt cache serves that history at a fraction of the input price, which is why a long session costs far less than its token count suggests, and why a session left idle past the cache's lifetime costs more when it picks up again.

What a developer costs a month

Anthropic's own figure, across enterprise deployments: about $13 per developer per active day and $150 to $250 per developer per month, with costs below $30 per active day for 90% of developers. The spread comes from a few habits more than from how much code people write.

• The model. Opus and Fable cost two to five times Sonnet per token, and Sonnet handles most coding work.
• Long sessions that are never cleared, which carry all their history into every request.
• Several sessions at once, subagents, scheduled tasks and agent teams: each runs its own requests, and Anthropic measures agent teams in plan mode at about seven times the tokens of a single session.
• Automation: Claude Code in CI or on a schedule runs whether anyone watches it or not.

Seats or the API: which costs less?

• Seats cost the same however much people use them. For someone who uses Claude Code every day, a seat usually costs less than the same work at API prices; the price is that they sometimes wait for a limit to reset.
• The API suits work that runs without a person, such as Claude Code in CI or on a schedule, where a seat's limits would get in the way.
• Seats waste money in two ways: seats nobody has opened in weeks, and people on Max or a premium seat who use a fraction of a smaller plan. The opposite is a cost too: people on Pro or a standard seat who run out every week and wait.
• Most teams mix them: seats for daily users, the API or usage credits for bursts and automation.

How to estimate your team's bill

• Count your developers by how often they use it: every day, every week, now and then.
• Price people on seats, at the plan their use needs, and automation on the API, at your pilot's numbers or Anthropic's average.
• Run a pilot first, as Anthropic suggests: a small group for a few weeks gives you a baseline that beats any average.
• Add the other tools people already pay for: ChatGPT for Codex, Cursor, GitHub Copilot.

• Estimate your team's AI coding spend

How to keep the bill in check

• See who uses what: by person, team and project, on which plan, and who keeps running out.
• Move seats to fit: down a plan where people use a fraction of theirs, up where they keep hitting limits, and cancel the ones nobody uses.
• Set budgets with alerts before the month ends, not after the invoice arrives.
• Share the habits that cost least: clear between unrelated tasks, Sonnet by default, Opus for the hard parts.

Claude Code and Codex spend in one place

Tokaware does the first three across Claude Code and Codex: every Claude and ChatGPT seat at the rates you pay, and what Anthropic, OpenAI, Cursor and GitHub bill through their admin APIs, by person, team and project, with budgets, forecasts, chargeback and which seats to move. An admin key brings in the last 30 days with nothing to install.

• Tokaware's pricing
• Track Claude Code usage across a team