AI cost · Anthropic
Claude Code cost: what you actually pay
Claude Code isn’t a separate seat price — it runs off your Claude subscription limits (Pro/Max/Team) or pay-as-you-go Anthropic API tokens. Cost hides in $100–$200 Max seats that sit idle and in Opus token burn with no per-repo attribution.
Public pricing
| Plan | Price | Billing | Includes |
|---|---|---|---|
| Pro | $20 / mo≈$17 annual; 5-hour rolling usage window — light Claude Code use. | monthly or annual | — |
| Max 5× | $100 / mo | monthly | 5× Pro capacity |
| Max 20× | $200 / mo | monthly | 20× Pro capacity |
| Team Standard | $25 / seat/mo≈$20 annual, min 5 seats. | monthly or annual | — |
| Team Premium | $125 / seat/mo≈$100 annual, min 5 seats. | monthly or annual | — |
| Enterprise | CustomCommonly ~$20/seat + metered API usage, billed annually. | custom | — |
Alternatively billed purely usage-based via the Anthropic API at model token rates (see the `anthropic_api` entry). Opus-heavy workflows dominate cost.
Last updated: 2026-07-22. Sourced from Claude Code pricing (2026) ↗, Anthropic pricing (official) ↗.
Where the real Claude Code cost hides
The sticker is the smallest part of the story. For Claude Code, these are the line items that quietly inflate the bill:
- Max ($100/$200) seats that sit mostly idle
- Opus token burn on tasks Sonnet could handle
- API-metered usage with no per-developer or per-repo attribution
What an audit finds on Claude Code
Teams buy $100–$200 Max seats for developers who rarely hit limits, or run fully API-metered with no attribution, so no one can see which repos or people drive the spend.
Worked example: Claude Code at 100 people
Take 100 people on Claude Code Team Premium at the published $125 per person per month. The seat count is a round number we picked; the price is the one in the table above.
- Monthly
- $12,500
- Annual
- $150,000
- Typically recoverable
- $15,000–$22,500
That last figure is 10 to 15% of seat and subscription spend — the band an audit typically surfaces across public benchmarks and our own sample data, before any metered overage on top. On Claude Code it usually arrives as seats nobody has opened in 60 days, which is the cheapest dollar in the stack to get back: nobody loses a capability they were using.
Arithmetic on the published prices above. Your mix will differ — run your own numbers.
Claude Code pricing FAQ
Is Claude Code free?
No — it requires a paid Claude plan (Pro/Max/Team) or Anthropic API credits; usage is bounded by the plan’s limits or billed per token via the API.
Max 5× vs Max 20×?
Max 5× is $100/mo and Max 20× is $200/mo; the difference is capacity, so idle high tiers waste money.
How is Claude Code billed on the API?
Per token at each model’s rate — Opus is the priciest, and Opus 4.7’s tokenizer emits ~35% more tokens for the same text.
What does Claude Code cost for a team of 100?
At the published $125 per person per month on Team Premium, 100 people is $12,500 a month, or $150,000 a year. An audit typically finds 10 to 15% of that recoverable — $15,000 to $22,500 a year — mostly as seats nobody has opened in 60 days.
Are these Claude Code prices current?
They were last checked against Claude Code pricing (2026) and Anthropic pricing (official) on 2026-07-22. Every price on this page carries that date and a link to its source, and a scheduled check fails our build when any figure goes more than 30 days unverified. Vendors do change pricing between checks — the source link is there so you can confirm before you sign.
What is Claude Code really costing you?
Spendassay connects your Claude Code spend to actual usage and engineering output — a proof-level AI cost report with the wasted dollars named. Neutral across every AI vendor.
Free · read-only · no card
Other coding assistants pricing
More: the complete AI cost management guide · every tool, with prices · cost per outcome · compare platforms · how we measure