GitHub Copilot Pricing 2026: Free, Pro, Pro+ and Business Plans

GitHub Copilot pricing in 2026 starts at $0 for the Free plan, then $10/month for Pro, $39/month for Pro+ and $100/month for the new Max tier. Teams pay $19 per seat per month on Business and $39 per seat on Enterprise. The big change is how usage works: since June 1, 2026, Copilot bills chat and agent work in “AI Credits” instead of premium requests, while code completions stay unlimited on paid plans and never touch your credits.

If you mainly want autocomplete in your editor, Pro at $10 is still one of the best value tools in software. If you lean on agent mode and frontier models all day, the credit allowance is what you are really buying, and that changes which plan makes sense. This guide walks through every plan, shows how AI Credits get used with real arithmetic, and tells you which tier to pick.

GitHub Copilot plans and prices at a glance

Here is the full lineup as listed on GitHub’s plans page and docs in September 2026. The “credits value” column is how GitHub itself presents each allowance, since one AI Credit equals one cent.

Extra credits on any plan cost $0.01 each. You can confirm the current numbers on the official Copilot plans page.

How AI Credits work since June 2026

On June 1, 2026, GitHub replaced the old premium request system with usage-based AI Credits. Instead of counting each chat message as one request, Copilot now meters the tokens each request uses and charges credits at that model’s published API rate. A token is a small chunk of text, roughly three quarters of a word; if the idea is new to you, our plain English guide to AI tokens explains it in a few minutes.

The practical effect is simple. Short questions to a cheap model cost almost nothing. Long agent sessions that read dozens of files and use a frontier model cost much more. GitHub explains the reasoning in its announcement of usage-based billing.

What never uses credits

Code completions (the grey ghost text that appears as you type) and Next Edit suggestions stay included on every plan and do not consume AI Credits. On Free you get 2,000 completions a month; on paid plans completions are not metered against your credit pool. For many developers, completions deliver most of the day-to-day value, which is why the $10 Pro plan still feels generous.

Base credits vs flex credits

Each individual plan splits its allowance into base credits and flex credits. Pro, for example, lists 1,000 base plus 500 flex for a total of 1,500. For budgeting, treat the total as your monthly allowance and check your usage dashboard to see how GitHub is drawing it down.

A worked example of credit usage

Because credits follow API rates, you can estimate costs with real token prices. Assume a model priced like Claude Sonnet 5.5 on Anthropic’s API, at $2 per million input tokens and $10 per million output tokens. (GitHub bills at each model’s published rate, so check the model list in your settings for the exact figure.)

• Quick chat question: 20,000 input tokens (your question plus some file context) and 2,000 output tokens. Input: 20,000 x $2 / 1,000,000 = $0.04. Output: 2,000 x $10 / 1,000,000 = $0.02. Total $0.06, or about 6 credits.
• Agent session: 400,000 input tokens (the agent reads many files over several steps) and 40,000 output tokens. Input: $0.80. Output: $0.40. Total $1.20, or about 120 credits.

On Pro’s 1,500 credits, that is roughly 250 quick questions, or about 12 substantial agent sessions, per month. On Pro+ with 7,000 credits, the same agent session fits about 58 times. Pick a pricier flagship model and those numbers shrink; pick a budget model and they grow a lot. Our guide to picking the cheapest AI model for each task applies directly here.

Each GitHub Copilot plan in detail

Copilot Free: $0

Free gives you 2,000 code completions per month plus “limited chat and agent usage.” GitHub lists the coding agent as available on Free, and individual plans include access to more than 30 models from providers such as Anthropic (Claude), OpenAI (GPT), Google (Gemini), xAI (Grok) and Kimi. No card is needed.

Who it is for: students waiting on verification, hobbyists, and anyone evaluating Copilot before paying. Who should skip it: anyone coding daily. 2,000 completions can disappear in a couple of busy weeks. For more no-cost options, see our list of the best free AI tools you can use today.

Copilot Pro: $10/month

Pro is the default pick for individual developers. You get completions without the Free cap, 1,500 AI Credits per month (shown as $15 in credits, so you receive more credit value than you pay), the full model picker and the coding agent.

Who it is for: professional developers who mostly use autocomplete and ask chat questions a few times a day. Who should skip it: people who run long agent tasks every day. They will likely buy overage or move up to Pro+.

Copilot Pro+: $39/month

Pro+ raises the allowance to 7,000 credits ($70 of value for $39) and adds third-party coding agent delegation, which lets you hand tasks to other vendors’ agents from inside GitHub. The value per dollar is the best of the individual tiers.

Who it is for: developers who use agent mode or frontier models for several hours a day. If you regularly spend $15 to $30 a month in overage on Pro, Pro+ is cheaper.

Copilot Max: $100/month

Max is the new top individual tier, with 20,000 credits ($200 of value). It targets people who treat Copilot’s agent as a full-time pair programmer, running multiple long sessions a day.

Who it is for: freelancers and consultants who bill by output and run agents constantly. Who should skip it: almost everyone else. Start on Pro+ and upgrade only if your dashboard shows you exhausting 7,000 credits by mid-month.

Copilot Business: $19 per seat per month

Business includes 1,900 credits per user per month, and the credits are pooled across the organization. Pooling matters: a lead engineer who runs heavy agent sessions can draw on credits that lighter users leave untouched. Admins get central controls over seats and usage. Our explainer on when AI team plans actually save money covers the seat math in more depth.

Copilot Enterprise: $39 per seat per month

Enterprise doubles the pooled allowance to 3,900 credits per user and adds the governance features larger companies need. See the Copilot plans documentation for the full feature matrix.

Legacy premium requests: what happened to the 300 limit

Older articles say Copilot Pro includes 300 premium requests per month, with overage at $0.04 per request. That no longer applies to new or monthly subscribers. Annual Pro and Pro+ subscribers who signed up before the change keep premium requests (300 on Pro, 1,500 on Pro+) until their term ends. After that, they drop to Free and can upgrade to a credit-based plan.

Hidden costs and how overage works

The sticker price is only part of the bill for heavy users. Keep these costs in mind:

• Overage at $0.01 per credit. A Pro user who burns 3,000 credits in a month pays $10 plus 1,500 x $0.01 = $15 in overage, for $25 total. At that point, Pro+ at $39 with 7,000 credits gives you far more headroom for only $14 more.
• Frontier model premiums. Credits follow API prices, so the most capable models burn credits several times faster than budget ones. Compare rates in our breakdowns of OpenAI API pricing and Claude API token costs.
• Big context. Agents that read your whole repository send a lot of input tokens. Tight, well-scoped tasks cost less. Our context window explainer shows why more context is not always better.
• Annual pricing. GitHub’s current pages do not show annual prices for Pro, Pro+ or Max. Historically Pro was $100/year, but confirm before assuming a discount exists.

How to save money on GitHub Copilot

1. Check free eligibility first. Verified students get the free Copilot Student plan. Verified teachers and maintainers of popular open source projects can get Copilot Pro free, with eligibility rechecked monthly. Details are in GitHub’s docs, and our roundup of AI student discounts lists similar offers.
2. Start on Pro and watch the dashboard. Run one full month on Pro and look at your credit usage before upgrading. Many developers never exceed 1,500.
3. Match the model to the task. Use a budget model for boilerplate and explanations. Reserve flagship models for architecture and tricky bugs.
4. Keep prompts and context tight. Point the agent at the specific files it needs. The techniques in our guide to reducing token usage apply to Copilot as well.
5. Upgrade when overage passes the gap. If overage on Pro regularly exceeds $29 (the difference to Pro+), move up. The same logic applies from Pro+ to Max at $61.

Which GitHub Copilot plan should you pick?

GitHub Copilot vs Cursor on price

Cursor Pro costs $20/month, double Copilot Pro, and also bills frontier models at API rates once included usage runs out. Copilot is cheaper for autocomplete-first developers and fits neatly if your code already lives on GitHub. Cursor tends to win for people who want an AI-first editor built around agents. Read our full Cursor vs GitHub Copilot comparison and the Cursor pricing breakdown before you decide, or browse the best AI coding assistants for 2026.

Frequently asked questions

Final verdict on GitHub Copilot pricing

GitHub Copilot remains excellent value in 2026. Pro at $10 is the right starting point for nearly every individual developer, because completions are unmetered and 1,500 credits cover regular chat. Move to Pro+ once agent mode becomes a daily habit, and consider Max only if your usage dashboard proves you need it. Teams should start on Business and use pooled credits to absorb uneven usage. Copilot pricing changed significantly this year and may change again, so confirm the current plans and credit amounts on GitHub’s official site before you subscribe.

Related reading:

• Cursor vs GitHub Copilot: which AI coding tool should you use?
• 10 best AI coding assistants for developers