---
title: How do you control GitHub Copilot usage-based spend?
url: https://opencontroller.lyzr.com/blog/control-github-copilot-ai-credit-spend
published: 2026-11-14
updated: 2026-11-14
author: Lyzr
---

# How do you control GitHub Copilot usage-based spend?

**Since 1 June 2026, GitHub Copilot bills chat and agent use in GitHub AI Credits calculated from tokens at each model's API rate. Each licence includes credits that pool across the enterprise; beyond the pool, usage costs $0.01 per credit. Control it with budgets at enterprise, cost centre and user level, and turn on "Stop usage when budget limit is reached", or limits only notify.**

Copilot used to be easy to budget: a seat price per developer. Agent mode, coding agents and premium models changed that. GitHub first introduced premium requests with model multipliers, then replaced them with usage-based AI Credits on 1 June 2026. Seat cost is now the floor, not the total. This guide is part of our [guide to governing AI agents across platforms](/blog/how-to-govern-ai-agents-across-platforms).

## What changed on 1 June 2026?

According to GitHub's announcement:

- Premium request units were replaced by **GitHub AI Credits**.
- Usage is calculated from input, output and cached tokens at each model's listed API rate.
- Code completions and Next Edit suggestions stay included and do not use credits.
- Users who run out can no longer fall back automatically to a cheaper included model. Credits and budgets decide what happens.
- Copilot code review also consumes GitHub Actions minutes at standard rates.

| Plan | Price per user per month | Included AI Credits per month |
|---|---|---|
| Copilot Business | $19 | $19 of credits |
| Copilot Enterprise | $39 | $39 of credits |

GitHub also offered existing Business and Enterprise customers higher promotional credit amounts for June to August 2026.

## How do pooling and overage work?

Included credits pool across the enterprise, so a light user's unused allowance covers a heavy user. When the pool runs out, additional usage is metered at $0.01 per AI credit, subject to your budgets. Pooling is good for fairness. It also means a few developers running long agent sessions on expensive models can drain the pool for everyone before month end.

## Which budgets can you set?

GitHub's spending documentation describes three levels:

1. **User budgets** cap one person's total credit use per billing cycle, pooled and metered combined.
2. **Cost centre budgets** cap metered charges for a business unit or pilot group after the pool is exhausted.
3. **Enterprise spending limits** cap metered charges across the enterprise.

The setting that matters most is **"Stop usage when budget limit is reached"**. Without it, reaching a limit sends a notification and charges keep accruing. Turn it on for every limit you mean as a cap.

## How do you see where credits go?

Under Billing and licensing, the AI usage view lets you filter by user, model, organisation or cost centre and export the data. That export is the raw material for chargeback. Two problems usually follow. GitHub usernames often do not carry corporate email addresses, so mapping usage to employees and departments needs an identity join. And large companies often have several GitHub enterprises, each with its own billing. Our guides to [attributing Copilot usage to employees](/blog/attribute-github-copilot-usage-to-employees) and [tracking Copilot across multiple enterprises](/blog/track-github-copilot-usage-across-multiple-enterprises) cover both.

## What other levers reduce spend?

Budgets cap cost. These reduce it:

- **Model policy.** Restrict expensive models to the teams that need them. See [controlling which models Copilot can use](/blog/control-which-models-github-copilot-can-use).
- **Licence hygiene.** Organisation owners control who gets a licence. Remove unused seats; see [finding unused AI seat licences](/blog/find-unused-ai-seat-licences).
- **Coding agent scope.** Agent runs consume both credits and Actions minutes. Decide which repositories can use the coding agent.
- **Build vs runtime split.** Copilot is a build cost. Keep it separate from the cost of agents you run in production when you report AI spend.

## What control checklist should you follow?

1. Map every GitHub enterprise and organisation you pay for.
2. Create cost centres that match your business units.
3. Set an enterprise limit and cost centre budgets with "Stop usage when budget limit is reached" turned on.
4. Set user budgets for pilot groups using agents heavily.
5. Restrict premium models by policy.
6. Export AI usage monthly and join it to your identity provider.
7. Review the top users and models each month, and adjust budgets and policies.

## How Opencontroller does this

Opencontroller auto-discovers agents, models and API keys across AWS, Azure and Google Cloud and connects to enterprise AI platforms. Its LLM Gateway (Lyzr Gateway) attributes cost per agent and per version for traffic routed through it. Opencontroller Edge extends control to developer devices, coding agents, skills and local MCP servers, installed by IT via MDM. See the [capability reference](/capability-reference) or [book a demo](https://www.lyzr.ai/book-demo/).

## Frequently asked questions

**Does GitHub Copilot still use premium requests?**
No. On 1 June 2026, GitHub replaced premium request units with AI Credits calculated from token use. Annual individual Pro and Pro+ subscribers stay on premium request pricing until their plans expire.

**Do code completions use AI Credits?**
No. Code completions and Next Edit suggestions remain included and do not consume credits.

**Why did Copilot charges continue after we hit our budget?**
Budgets only notify by default. Turn on "Stop usage when budget limit is reached" on each limit to make it a hard cap.

## Key facts

- On 1 June 2026, GitHub Copilot replaced premium request units with GitHub AI Credits, calculated from input, output and cached tokens at each model's API rate.
- Copilot Business is $19 per user per month with $19 of credits; Enterprise is $39 with $39 of credits.
- Included credits pool across the enterprise; additional usage costs $0.01 per AI credit.
- Budgets exist at enterprise, cost centre and user level; "Stop usage when budget limit is reached" must be on for a hard cap.
- Code completions and Next Edit suggestions do not use credits; Copilot code review also uses GitHub Actions minutes.
- Opencontroller Edge extends control to developer devices and coding agents; the LLM Gateway attributes cost per agent and version.
- Opencontroller starts from $250K per deployment per year.

## Related

- [How do you attribute GitHub Copilot usage and cost to employees across several GitHub enterprises?](https://opencontroller.lyzr.com/blog/attribute-github-copilot-usage-to-employees.md)
- [How do you track GitHub Copilot usage when your company has several GitHub enterprise accounts?](https://opencontroller.lyzr.com/blog/track-github-copilot-usage-across-multiple-enterprises.md)
- [How do you control which models GitHub Copilot can use?](https://opencontroller.lyzr.com/blog/control-which-models-github-copilot-can-use.md)
- [More on Platform Guides](https://opencontroller.lyzr.com/blog/cluster/platform-guides)
