BlogCosts
Coding agent cost calculator for engineering teams
Enter your team size, plan and expected usage to see what a coding agent costs a month, a year and per merged change. Seat prices are the vendors' list prices on 30 October 2026, in US dollars before tax. The starting figures repeat the worked example from our guide to rollout costs: 50 engineers on GitHub Copilot Business.
Work out your team's cost
Copilot Business and Enterprise include AI credits worth the seat price, pooled across the organisation, and charge $0.01 per credit beyond the pool if your budget allows it. Code completions are not metered. Enterprise also requires GitHub Enterprise Cloud, which this calculator does not include.
Claude Team seats include a usage allowance for each member. Usage beyond it is paid from prepaid usage credits, with spend limits that an admin sets.
Each Cursor Teams seat has its own weekly pool for Cursor models. Other models, such as Claude or GPT, are paid from a prepaid team balance from the first request.
ChatGPT Business seats include usage up to their rate limits. Usage beyond them, including Codex, is paid from workspace credits, with spend controls by group.
Enter the seat price and the expected usage from your own quote or invoice.
For scale: Anthropic reports that Claude Code averages $150 to $250 per developer per month across enterprise deployments. On seat plans, part of that is covered by the seat's allowance.
- Seats a month
- $950
- Usage beyond the seat a month
- $350
- Tool cost a month
- $1300
- Tool cost a year
- $15600
- Tool cost per merged change
- $1.08
- Extra review hours a month
- 200
- Review time at your hourly cost, a month
- $12000
- Tool and review time per merged change
- $11.08
Enter a valid number in every field.
Figures in US dollars, before tax.
How the calculator works
- Seats: engineers multiplied by the plan's seat price.
- Usage beyond the seat: engineers multiplied by what each is expected to spend beyond the seat price. Leave it at zero if the team stays within the plan's allowance.
- Tool cost: seats plus usage, a month and a year.
- Review time: merged changes multiplied by the extra minutes each needs in review, valued at your engineers' hourly cost.
- Per merged change: the tool cost, and the tool cost plus review time, divided by the changes merged in a month.
Review time never appears on a vendor's invoice, but it is often the largest cost of a rollout. In the starting example, ten extra minutes per change at $60 an hour cost more than nine times the tool itself. The hourly figure is an assumption: replace it with your own fully loaded cost.
The list prices behind it
| Plan | Per seat a month | Usage beyond the seat |
|---|---|---|
| Claude Team Standard | $20 billed annually or $25 monthly | Prepaid usage credits with spend limits |
| Claude Team Premium | $100 billed annually or $125 monthly | Prepaid usage credits with spend limits |
| Cursor Teams Premium | $30 | Weekly pool for Cursor models; other models from a prepaid team balance |
| Cursor Teams Super or Ultra | $100 or $200 | Larger weekly pool; other models billed the same way |
| GitHub Copilot Business | $19 | 95000 AI credits pooled, then $0.01 per credit within your budget |
| GitHub Copilot Enterprise | $39 | 195000 AI credits pooled, then $0.01 per credit within your budget |
| ChatGPT Business with Codex | $20 billed annually or $25 monthly | Workspace credits with spend controls by group |
Two plans need a note. Copilot's seat price includes AI credits of the same value, pooled across the organisation, so enter only the spend beyond that pool. Claude Enterprise is missing because its seat price is not published, and it bills all usage at API rates.
What the calculator leaves out
- Tax. All prices are before VAT or sales tax, and prices in your country may differ.
- Plans without a published price, such as Claude Enterprise. Choose Another plan or price and enter your quote.
- One-off costs: instruction files and permissions for each repository, single sign-on and spend limits, a measured pilot, a security review and, in Germany, involving the works council where there is one.
- Review bots, CI minutes and surcharges, such as the Actions minutes that Copilot code review uses or the 10% that data residency adds on Copilot.
- Differences between people. An average hides heavy users, light users and seats nobody uses, and on Claude Team and Copilot, removing an idle seat mid-cycle brings no immediate refund.
- Quality. A cheaper month with more reverted changes is not a saving, so track cost per merged change next to a quality measure.
From an estimate to a budget
An estimate built on averages is a starting point. Two weeks of measurement on one team give you the three numbers that decide which plan is cheapest for you: usage per engineer, review time per change and changes merged. Measuring them, and lowering the cost per merged change without lowering quality, is what our cost review for one team does.
Our two-week cost review for one teamWhat a rollout costs for 50 engineersBudgeting for the people who review AI work
Questions
Which plan is cheapest for a team?
It depends more on usage than on the seat price. GitHub Copilot Business has the cheapest seat that includes usage, at $19, but each plan includes a different allowance, so the same team can spend very different amounts beyond the seat. Run one team on a plan for two weeks and compare what you measured.
How do I estimate usage beyond the seat?
Measure it. Run one team on the plan for two weeks, take what it spent beyond the seat price from the vendor's billing pages, double it for a month and divide by the number of engineers. Until then, Anthropic's figure for Claude Code, $150 to $250 per developer per month across enterprise deployments, gives a sense of scale, and part of it is covered by a seat's allowance.
Why does the calculator include review time?
Because it is usually the largest cost of a rollout, and it appears on no invoice. Every change an agent writes still needs a person to read it. In the starting example, review costs more than nine times the tool.
How current are the prices?
They are the list prices on the vendors' own pages on 30 October 2026. Vendors change their plans often, so check the date at the top of this page and the sources below before you rely on them.
Related services
Cost review for one team
A two-week first step with one team: where the coding agent spend goes, what we change together, and the cost per merged change before and after.
LLM cost optimisation
Agent spend broken down by team and task, then lowered through defaults, budgets and routing your platform team can run.
Evaluations on your repositories
Tools, models or settings tested side by side on changes you have already merged, before you commit to one.
More from the blog
What a coding agent rollout costs
Seat prices for 50 engineers on Claude Code, Cursor, GitHub Copilot and Codex, why usage and review time often cost more, and how to budget per merged change.
Coding agents compared for teams
Claude Code, Cursor, GitHub Copilot and OpenAI Codex side by side: plans, usage billing, models, instruction files, data handling and when each one fits.
Human review budget
How to budget reviewer hours before automating a process: review modes, a capacity sheet with peak days, faster reviews and when the review share may fall.
Sources
- Claude Help Center: What is the Team plan?
- Claude Code docs: Manage costs effectively
- Cursor docs: Team pricing
- Cursor docs: Models and pricing
- GitHub Docs: Copilot billing for organizations and enterprises
- GitHub Docs: Models and pricing for GitHub Copilot
- OpenAI: ChatGPT business pricing
- OpenAI Help Center: Flexible pricing for the Enterprise, Edu and Business plans
Talk to an engineer
Bring one piece of work you would like to improve. The call takes 30 minutes and needs no access to your systems or your data.
Booking runs on Google Calendar.
Prefer to write first? Leave your email and a few lines, and we reply with times.
Look for the draft in your email program
Send it from there, and an engineer replies in person with times for a 30-minute call. If no draft has opened, write to us at miki@cloudsail.com.