Budget AI coding tools by piloting with a small group, measuring real usage for a few weeks, and using public averages only as a scale check. Anthropic's documentation puts the enterprise average around $150 to $250 per developer per month, which would be $3,000 to $5,000 a month for 20 people.
Start from published scale, not a promise
Anthropic's Claude Code cost page says that across enterprise deployments the average is around $13 per developer per active day and $150 to $250 per developer per month, with costs below $30 per active day for 90 percent of users. It also says per developer cost varies widely with model choice, codebase size and usage patterns such as running multiple instances or automation, and recommends starting with a small pilot to set a baseline.
For 20 developers, $150 to $250 each gives $3,000 to $5,000 a month. Treat that as the rough size of the question, not the answer.
A four week pilot
Week 1: choose 3 to 5 developers across different work (new features, bug fixes, tests). Set a workspace spend limit so the pilot cannot run away.
Weeks 2 and 3: let them work normally. Record spend per person and per day, which models they used, and when they hit limits.
Week 4: look at the spread, not only the average. Ask what the heaviest user was doing and whether it was worth it.
What to measure
- Cost per active day, since not everyone codes daily.
- Share of spend by model.
- Cache hit behavior and long sessions, which the usage view can flag.
- Work shipped, such as merged pull requests, so cost has a denominator.
Size the rate limits
Anthropic's page gives per user token per minute recommendations by team size, for organizations that request limits. For 20 to 50 users it lists 50k to 75k TPM and 1.25 to 1.75 RPM per user, for 50 to 100 users it lists 25k to 35k TPM and 0.62 to 0.87 RPM per user. The page's own example: 200 users at 20k TPM each is 4 million TPM in total. These apply at organization level, so individuals can temporarily exceed their share when others are idle. Plan extra for live training sessions with many people at once.
Choose who pays how
Anthropic lists three setups: Teams or Enterprise seats with spend limits in admin settings, Console API billing with workspace limits, or a cloud provider with your cloud's budget controls. See subscription or API and track AI spend per team.
Set the cap before the rollout
Decide the monthly number your finance team accepts, put a limit just under it, and decide what happens at 80 percent. For wider guidance on enterprise budgets, see enterprise LLM budget management patterns.
Quick check
Is $13 a day what I should expect?
It is an average Anthropic reports across enterprise deployments, not a forecast. Your pilot tells you more.
Who should see per developer numbers?
Decide that with engineering leadership first. Usage data is useful for support and cost control, and it can feel like surveillance if it is used badly.
Sources
Last reviewed: 19 October 2026.
