GitHub Copilot Business costs $19 per granted seat per month and Copilot Enterprise costs $39, so a 10 person team runs $190 or $390 a month, a 50 person team $950 or $1,950, and a 200 person team $3,800 or $7,800. Annually that is $11,400 for 50 Business seats and $93,600 for 200 Enterprise seats. Those numbers are only the floor, though. Since June 1, 2026 every seat also carries a pooled allowance of GitHub AI credits, 1,900 on Business and 3,900 on Enterprise, and anything your team burns past the shared pool bills at $0.01 per credit on top of the seat price.
The seat arithmetic is simple enough that nobody gets it wrong. What catches finance teams out is the second line on the invoice, and a promotional allowance that has now expired. Here is the whole picture.
What GitHub Copilot costs per month by team size
You are billed per granted seat, not per person in the organization, so licenses you have not assigned cost nothing. That makes the base calculation clean:
| Team size | Business at $19 | Enterprise at $39 | Business per year | Enterprise per year |
|---|---|---|---|---|
| 10 developers | $190 / mo | $390 / mo | $2,280 | $4,680 |
| 25 developers | $475 / mo | $975 / mo | $5,700 | $11,700 |
| 50 developers | $950 / mo | $1,950 / mo | $11,400 | $23,400 |
| 100 developers | $1,900 / mo | $3,900 / mo | $22,800 | $46,800 |
| 200 developers | $3,800 / mo | $7,800 / mo | $45,600 | $93,600 |
For most companies these are not frightening numbers. A 50 seat Business plan at $11,400 a year is less than a fifth of one mid-level engineer's fully loaded cost, and cheaper than bringing in vetted contract capacity to work through the same backlog. The question is almost never whether $19 a seat is affordable. It is whether $19 is the number you will actually pay.
The credit pool is the line that moves
Every Copilot seat contributes its AI credit allowance to one shared organization pool. This is the mechanic most cost write-ups skip, and it works in your favor. Fifty Business seats is a single pool of 95,000 credits a month, not fifty separate buckets of 1,900. Your two engineers who run the agent constantly draw from the same balance as the fifteen who use Copilot as autocomplete and barely touch it. You do not have to size the plan around your heaviest user.
| Team size | Business pool | Enterprise pool |
|---|---|---|
| 10 developers | 19,000 credits | 39,000 credits |
| 50 developers | 95,000 credits | 195,000 credits |
| 200 developers | 380,000 credits | 780,000 credits |
Two things decide whether that pool holds. The first is good news: code completions and next edit suggestions are not billed in credits at all and stay unlimited on every paid plan. If your team mostly uses Copilot to finish lines of code, the pool barely moves and a seat behaves like a flat fee. Credits are consumed by chat, agent mode, the cloud coding agent and AI code review.
The second is that credits are drawn against published per-model API rates, counting input, output and cached tokens. So a long agent session on a frontier model can cost many multiples of a short chat on a small one, and no allowance converts cleanly into a number of prompts. We wrote up the mechanics separately in what GitHub AI credits are.
The promotional allowance has ended
Every allowance quoted above is the standard one, which is what you should be budgeting against today. It is still worth knowing what came before it. GitHub ran a temporary promotional allowance for existing organization customers from June through August 2026: Business seats received $30 of monthly credits instead of $19, and Enterprise seats $70 instead of $39. In credit terms that was 3,000 a seat rather than 1,900, and 7,000 a seat rather than 3,900. It ended with August 2026 and was not renewed.
Work that through a 50 person team and the drop is not subtle:
| Plan, 50 seats | Pool through Aug 2026 | Pool from Sep 2026 | Change |
|---|---|---|---|
| Copilot Business | 150,000 credits | 95,000 credits | Down about 37% |
| Copilot Enterprise | 350,000 credits | 195,000 credits | Down about 44% |
Same seat price, considerably less included usage. A 50 seat Business org that settles at 120,000 credits a month sat comfortably inside the promotional pool and is now 25,000 credits over the standard one, which adds $250 to a $950 bill. That is a 26 percent increase that a summer forecast would not have shown. If your numbers were built before September 2026, rebuild them on the standard allowance before you sign off on next quarter.
What happens if you go over
Usage beyond the pool bills at $0.01 per AI credit. Credit where it is due: that is exactly what your included credits are worth, so GitHub does not mark up overage the way several competitors do. Admins can also set budgets at the user, cost center and enterprise level, which means going over is a decision you make rather than a surprise you discover. You can cap spend and let work stop at the ceiling, or allow it to continue at a rate you can forecast.
That combination, pooled allowances plus unmarked-up overage plus real budget controls, makes Copilot one of the more honest meters in this category. It is worth saying plainly, because the trend across the rest of the market runs the other way. Our comparison of flat and credit-metered AI coding tool pricing tracks who charges what once you cross the line.
Should you buy Business or Enterprise?
Start on Business. The extra $20 a seat only earns its place where the larger credit pool actually gets consumed, and because credits pool organization-wide you can often absorb one heavy team inside a Business allowance rather than upgrading everyone. There is also a requirement that settles the question for a lot of companies before price enters it: Copilot Enterprise runs on GitHub Enterprise Cloud. If you are on a standard GitHub organization, Business is the only option, and the real cost of moving up is not $20 a seat, it is $20 a seat plus a platform contract.
One more piece of arithmetic worth carrying into the decision. At the published $0.01 credit rate, Copilot Pro+ costs $39 a month and includes 7,000 credits, while a Copilot Enterprise seat costs the same $39 and includes 3,900. Same price, 79 percent more usage on the individual plan. That is not a pricing error, it is the shape of the product: an org seat prices license management, policy control, content exclusion, budgets and audit. You are buying governance, and the AI usage comes at cost. The full breakdown lives on our GitHub Copilot Business pricing page, and the plan-by-plan differences in Copilot Business vs Enterprise.
How to budget for it honestly
Three steps that take about twenty minutes and save an awkward conversation later.
First, pull your current credit consumption from the organization billing dashboard and divide it by the standard allowance for your seat count, so you know your true utilization rather than the flattering summer version. Second, multiply your monthly credit burn by $0.01 and compare it against that standard pool; the gap is your overage in dollars. Third, set a budget at the enterprise level slightly above that figure rather than leaving spend uncapped, so an experiment with the cloud agent cannot quietly triple a month.
If the forecast comes back uncomfortable, the answer is usually not a bigger plan. It is deciding which work genuinely needs an agent and which is fine on completions, which are free. The individual tiers are covered in our GitHub Copilot pricing guide if you are also weighing what to give contractors or a small team outside the org.
The alternative worth pricing alongside it
Copilot Business is a strong buy for a company standardizing on GitHub, and we are not going to pretend otherwise. Unlimited unmetered completions on every paid seat is a genuine advantage, and the admin surface is the most mature in the category. Two things it does not solve: your code has to be on GitHub, and what you get is a faster editor rather than work you can hand off and come back to.
That is the gap we build for. Agentcode is $29 a month flat, runs on GitHub and GitLab, and works asynchronously: you describe a task, the agent plans it, edits the code, runs your tests and opens a pull request you review. It never merges on its own. There is no credit pool to watch, no seat math to redo when an allowance changes, and a busy month costs exactly what a quiet one does. Most teams that adopt it keep Copilot for completions and use an async agent for the tasks nobody has time to start.
Keep reading
- GitHub Copilot Enterprise pricing against Copilot Pro+: which $39 seat actually buys more AI credits
- Claude Code Bedrock pricing: what an AWS team actually pays, and what the AWS bill will not tell you
- Codex VSCode extension pricing: which plan pays for Codex in VS Code, and what each one costs
- CodeRabbit pricing against GitHub Copilot code review and Greptile: what a team of ten actually pays