D
DevToolsReview

GitHub Copilot Pricing 2026: Free, Pro $10, Pro+ $39, Business $19

Every GitHub Copilot plan in one table: Free, Pro $10, Pro+ $39, Max $100, Business $19/user, Enterprise $39. What AI credits buy, and team costs.

DR

DevTools Review

· Updated November 7, 2026
GitHub Copilot

Get told when a price changes

We re-verify all 12 tools. You hear from us only when something actually moved.

Buying for a team? Price your team's bill on every plan, free. Takes under a minute.

GitHub Copilot costs $0 on Free, $10 a month on Pro, $39 on Pro+, $100 on Max, $19 per user on Business and $39 per user on Enterprise. Those sticker prices haven’t moved since July. What changed is what the $10 buys. On 1 June 2026 GitHub retired “premium requests” and moved every plan to GitHub AI Credits — a monthly dollar allowance that drains at each model’s per-token API rate.

That makes most Copilot pricing pages wrong in a way that matters. “300 premium requests” and “unlimited chat” no longer describe any plan you can buy. Here’s what each tier costs and includes, re-verified against GitHub’s plans page and billing docs on 7 November 2026. For features and day-to-day performance, see our full Copilot review.

Quick Summary

PlanPriceMonthly AI creditsBest for
Free$0Small allowance, 2,000 completionsTrying Copilot
Student$0 (verified students)Allowance, unlimited completionsStudents
Pro$10/mo1,500 (1,000 base + 500 flex)Most individual developers
Pro+$39/mo7,000 (3,900 base + 3,100 flex)Daily agent-mode users
Max$100/mo20,000 (10,000 base + 10,000 flex)Heavy, sustained agent use
Business$19/user/mo1,900 per user, pooledTeams needing admin + IP indemnity
Enterprise$39/user/mo3,900 per user, pooledGitHub Enterprise Cloud orgs

One AI credit is one US cent, so Pro’s 1,500 credits are $15 of model usage at GitHub’s published per-token rates. For individual developers, Pro at $10/month is still the clear winner — it’s the only mainstream plan that hands you more usage than you pay for.

Official GitHub Copilot site (github.com)

What Changed on 1 June 2026

GitHub announced the switch on 27 April 2026. Three things are different now.

1. Requests became tokens

Under the old system, every chat message or agent step cost one “premium request” (times a model multiplier). Now an interaction costs whatever its tokens cost — input, output, and cached tokens, priced per model. A quick question to a lightweight model costs a fraction of a credit. A long cloud-agent session on a frontier model across many files can cost dollars.

2. Your allowance is “base + flex”

Each paid individual plan includes base credits equal to its price (Pro: 1,000 = $10) plus a flex allotment on top (Pro: 500). Base is fixed. Flex is explicitly variable — GitHub’s docs say it is “designed to adapt as the economics of AI evolve.” Read that as: the bonus can shrink.

Credits reset at 00:00 UTC on the 1st of each calendar month regardless of your billing date, and unused credits don’t roll over.

3. Business and Enterprise lost their promo credits

Business and Enterprise seats got extra promotional credits from June to August. Those ended on 31 August 2026: included usage fell from $30 to $19 per Business seat and from $70 to $39 per Enterprise seat, with no change to the seat price. The team allowance is pooled across the billing entity, so light users subsidise heavy ones. We worked through what that does to a team invoice in Copilot’s September credit drop.

Annual subscribers who bought before June stay on legacy premium-request billing until their plan expires, then move to credits.

What Do 1,500 Credits Actually Buy?

This is the question every other pricing page skips. GitHub publishes per-model token rates, so you can estimate. These are our illustrative estimates, not GitHub figures:

  • A quick chat question on GPT-5 mini (~5K tokens in, ~1K out): about $0.003 — a third of a credit. Pro’s allowance covers thousands of these.
  • A substantial agent-mode task on Claude Sonnet 4.6 ($3 per million input, $15 per million output; assume ~200K input tokens, mostly cached, and ~20K output): roughly $0.50–$0.90. Pro covers about 15–30 tasks like that a month.
  • The same task on Claude Opus 5 ($5 in / $25 out): roughly $0.80–$1.50 each, so ten to twenty a month on Pro.

The pattern: Pro is generous for completions and chat, tight for daily agent work on frontier models. If you run agent mode several times a day, you’ll either buy extra credits at the same per-token rates or move to Pro+.

Two ways to stretch the allowance: switch heavy tasks to lighter models, and use auto model selection, which gives paid plans a 10% discount on model costs.

All Plans in Detail

Free ($0)

  • 2,000 code completions per month and a small allowance of AI credits for chat and agent mode
  • Auto model selection only — you can’t pick the model
  • Copilot CLI and agent mode are included; code review and the cloud agent are not
  • Works in VS Code, Visual Studio, JetBrains, Eclipse, Xcode and on github.com

Who it’s for: Evaluating Copilot, or a few hours of coding a week. Working developers will hit the completion cap within days.

Student ($0, verified students)

New since the credit switch. Verified students get unlimited code completions, an AI credit allowance, and auto model selection. Third-party coding agents (Claude, Codex delegation) are excluded.

This replaces the old “students get Pro free” deal. Verified teachers and maintainers of popular open-source projects may still be eligible for free Pro — check before paying.

Pro ($10/month)

What you get:

  • Unlimited code completions and next-edit suggestions — not billed in credits
  • 1,500 AI credits a month for chat, agent mode, Copilot CLI, Spaces, Spark and the cloud agent
  • Model choice across Anthropic, OpenAI, Google, xAI and Microsoft models, including Claude Sonnet 4.6/5, Claude Opus 5 and GPT-5.5
  • Cloud agent — assign an issue and Copilot opens a pull request
  • Code review on pull requests and in the editor
  • Buy extra credits when you run out, at the same per-token rates

Who it’s for: Every individual developer who codes regularly. Even if you never touch agent mode, unlimited completions alone justify $10.

Pro+ ($39/month)

Pro+ rose from $20 to $39 on 6 July 2026 — a 95% increase that many comparison pages still haven’t caught. It now includes 7,000 credits (3,900 base + 3,100 flex), about $70 of usage for $39.

The one feature-level difference from Pro: Pro+ can delegate tasks to third-party coding agents like Claude by Anthropic and OpenAI Codex (preview).

Who it’s for: Developers who run agent mode or the cloud agent every day and keep running out of credits on Pro.

Max ($100/month)

20,000 credits (10,000 base + 10,000 flex), about $200 of usage, plus priority access to premium models. GitHub positions it for “sustained, high-volume” agent use.

Who it’s for: People running long cloud-agent sessions daily. If you’re regularly buying more than ~$60 of top-up credits on Pro+, Max is cheaper.

Business ($19/user/month)

Everything in Pro, plus organisation controls:

  • 1,900 credits per user, pooled across the organisation, with admin spending budgets
  • Policy management — enable or disable features and models org-wide
  • IP indemnity for Copilot-generated code (with the public-code filter on)
  • Content exclusions — keep sensitive files and repos out of context
  • Audit logs, centralised seat management, and proxy support

Two cost notes for admins. Since 1 June, Copilot code review also consumes GitHub Actions minutes. And GitHub warns orgs that never enabled the old “premium request paid usage” policy to set budgets now, because overage is metered.

Enterprise ($39/user/month)

Everything in Business, plus 3,900 credits per user (pooled) and priority access to new models and features. Requires GitHub Enterprise Cloud — Copilot isn’t available on GitHub Enterprise Server.

Who it’s for: Large organisations already on GitHub Enterprise Cloud. Most teams under ~50 developers are better served by Business.

Pricing Comparison Table

FeatureFreePro ($10)Pro+ ($39)Max ($100)Business ($19/user)Enterprise ($39/user)
Monthly AI creditsSmall allowance1,5007,00020,0001,900 (pooled)3,900 (pooled)
Code completions2,000/moUnlimitedUnlimitedUnlimitedUnlimitedUnlimited
Choose your modelAuto onlyYesYesYes + priorityYesYes + priority
Agent mode & CLIYesYesYesYesYesYes
Cloud agent (issue → PR)NoYesYesYesYesYes
Code reviewNoYesYesYesYesYes
Third-party agents (Claude, Codex)NoNoYesYesPer org policyPer org policy
Buy extra creditsNoYesYesYesAdmin budgetAdmin budget
IP indemnity & policy controlsNoNoNoNoYesYes

At a Glance

How much does GitHub Copilot cost in 2026?
According to DevTools Review’s pricing verification on 7 November 2026, Copilot Pro costs $10/month, Pro+ is $39/month, Max is $100/month, Business is $19/user/month and Enterprise is $39/user/month. Copilot Free and Copilot Student cost nothing.
Does GitHub Copilot still use premium requests?
No. Since 1 June 2026 Copilot bills usage in GitHub AI Credits (1 credit = $0.01), charged per token at each model’s rate. Pro includes 1,500 credits a month, Pro+ 7,000 and Max 20,000. Only pre-June annual subscribers remain on premium requests until renewal.
Is GitHub Copilot cheaper than Cursor?
Yes, by a wider margin than before. Copilot Pro is $10/month. Cursor retired its $20 Pro plan in October 2026; its working tier is now Premium at $30/month, and its $8 Plus plan is limited to Grok models.
Can you use GitHub Copilot for free?
Yes. Copilot Free gives 2,000 completions a month and a small credit allowance, with auto model selection only. Verified students get Copilot Student free, with unlimited completions.
What changed for Copilot Business and Enterprise in September 2026?
The seat prices held at $19 and $39, but included AI credits fell from $30 to $19 per Business seat and from $70 to $39 per Enterprise seat when promotional credits ended on 31 August 2026.

What GitHub Copilot users report

Read Nov 7, 2026 on GitHub Community and Hacker News. These are individual, self-selected posts, not a survey, and unhappy users post more often. This round doesn't include Reddit.

  • Credits can vanish quickly

    Complaint

    This is the loudest theme since usage-based billing began on 1 June. GitHub's announcement drew 590 comments, and users describe a few prompts taking a large share of a month's credits, often on frontier models.

    GitHub announcement thread, May 27, 2026 · User complaint thread, Jun 11, 2026

  • Costs are hard to predict

    Complaint

    Users ask for a per-prompt cost breakdown because similar work can cost different amounts. One reports being billed for a different model than they picked; staff replied that some background features use other models.

    GitHub announcement thread, May 27, 2026

  • Some heavy users say it pays off

    Praise

    One developer reports spending about $500 in credits, mostly on Opus, and calls it far cheaper than the team time it replaced. Others in GitHub's thread plan to stay on the $10 plan unchanged.

    Hacker News comment, Oct 23, 2026 · GitHub announcement thread, May 27, 2026

  • Value compared with rivals

    Complaint

    One user who is required to use Copilot at work says it went from fantastic to poor value after the billing change, and believes the same money buys far more usage in Codex. That's their estimate, not a measurement.

    Hacker News comment, Sep 26, 2026

Hidden Costs and Gotchas

Usage is now metered. The old “flat rate, no surprises” story is over. Once your credits run out you either stop or pay per token. Set a budget in your billing settings before you discover this through a bill.

Flex credits can shrink. A third of Pro’s allowance (500 of 1,500) and nearly half of Pro+‘s is “flex”, which GitHub says may change over time. Judge each plan by its base credits; treat flex as a bonus.

Agent sessions are the expensive part. A single long cloud-agent run on a frontier model can use more credits than a week of chat. Use cheaper models for routine work and set session limits for agent tasks.

Credits reset on the 1st, not your billing date. If you subscribe on the 25th, your first allowance lasts six days.

No top-ups on mobile subscriptions. If you subscribed through GitHub Mobile on iOS or Android, you can’t buy extra credits.

Code review costs Actions minutes for teams. Business and Enterprise orgs pay for review twice: credits plus Actions minutes, since 1 June 2026.

Who Should Pick Which Plan?

Choose Free if you:

  • Code a few hours a week on side projects
  • Want to try Copilot before paying

Choose Student if you:

  • Are a verified student — never pay for Copilot as a student

Choose Pro if you:

  • Write code daily and mostly want completions and chat
  • Use agent mode a few times a day, not constantly
  • Are an individual developer, freelancer, or indie hacker

Choose Pro+ if you:

  • Run agent mode or the cloud agent all day
  • Want to delegate tasks to Claude or Codex agents from GitHub
  • Keep buying top-up credits on Pro

Choose Business if you:

  • Manage a team that needs central billing, policies and spending budgets
  • Need IP indemnity or content exclusions

Choose Enterprise if you:

  • Run a large org on GitHub Enterprise Cloud that needs the bigger pooled allowance

Is GitHub Copilot Worth $10/Month?

Yes — still the easiest “yes” in AI coding tools, with one new caveat.

Unlimited completions alone are worth $10 to anyone who codes daily. On top of that, Pro gives you $15 of model usage for $10. No other mainstream tool hands you more included usage than you pay for.

The caveat is agent mode. If you’ve adopted agentic coding as your main workflow, 1,500 credits won’t last the month on frontier models. Then the honest comparison is Pro+ ($39 for ~$70 of usage) against Claude Code at $20 or Cursor Premium at $30.

Where Copilot wins:

  • Works in every major editor — VS Code, Visual Studio, JetBrains, Eclipse, Xcode, Neovim and github.com. No editor switch.
  • GitHub integration — assign an issue, get a pull request; code review sits where your PRs already are.
  • Model choice — Anthropic, OpenAI, Google and xAI models in one subscription, billed at published rates.
  • Price floor — $10 is still the cheapest paid tier among the major tools.

Where Copilot falls short:

  • Metered agents — heavy agent users now pay per token, like everyone else.
  • Deep codebase context — Cursor’s project indexing is still stronger for large refactors.
  • Free tier is narrow — 2,000 completions and auto-only models.

See how it stacks up in our Cursor vs Copilot comparison, Copilot vs Claude Code, and the full AI coding tools pricing comparison.

GitHub Copilot vs. the Competition

Copilot Pro ($10) vs. Cursor Premium ($30): Cursor repriced in October 2026 — the $20 Pro plan is gone, and third-party models like Claude and GPT bill separately at API rates. Copilot is a third of the price and includes those models in its credits. Cursor still leads on codebase indexing and multi-file editing. See Cursor pricing.

Copilot Pro ($10) vs. Windsurf Pro ($20): Windsurf (now under Devin) costs twice as much and runs only in its own editor. Copilot wins on price and editor support unless you specifically want Windsurf’s Cascade workflow. See Windsurf pricing.

Copilot Pro+ ($39) vs. Claude Code Pro ($20): For terminal-first agentic work on Claude models, Claude Code’s $20 plan is the cheaper way in. Copilot Pro+ makes sense if you want agents inside GitHub and your editor, with multiple model vendors.

Copilot Business ($19/user) vs. Windsurf Teams ($80 base + $40/user): Copilot Business is far cheaper per seat and includes IP indemnity. For most teams already on GitHub, it’s the default organisational choice.

FAQ

How much does GitHub Copilot cost?

As of 7 November 2026: Copilot Pro is $10/month, Pro+ is $39/month (raised from $20 on 6 July 2026), Max is $100/month, Business is $19/user/month and Enterprise is $39/user/month. Copilot Free and Copilot Student (for verified students) are free.

What are GitHub AI Credits?

Since 1 June 2026, Copilot measures usage in GitHub AI Credits, where 1 credit equals $0.01. Chat, agent mode, Copilot CLI, Spaces, Spark and the cloud agent consume credits at each model’s per-token rate. Pro includes 1,500 credits a month, Pro+ 7,000 and Max 20,000. Code completions are not billed in credits.

Does Copilot still have premium requests?

No. Premium requests were retired on 1 June 2026 and replaced by AI credits. Only annual subscribers who bought before June remain on premium-request billing, until their plan expires.

Is GitHub Copilot free?

Yes. Copilot Free includes 2,000 code completions a month and a small AI credit allowance with auto model selection. Verified students get Copilot Student free with unlimited completions. See our guide to the best free AI coding tools for alternatives.

What happens when I run out of Copilot credits?

On Pro, Pro+ and Max you can buy more credits at the same per-token rates, or upgrade your plan. Code completions keep working because they don’t use credits. On Business and Enterprise you can request more budget from your administrator.

Is Copilot Pro or Pro+ better value?

Pro gives 1,500 credits for $10; Pro+ gives 7,000 for $39 and adds delegation to third-party agents like Claude and Codex. If you regularly buy more than about $30 of extra credits on Pro, Pro+ is cheaper.

Is GitHub Copilot cheaper than Cursor?

Yes. Copilot Pro costs $10/month. Cursor retired its $20 Pro plan in October 2026; its main tier is now Premium at $30/month, and third-party models on Cursor bill separately at API rates.

Can I use Copilot in JetBrains?

Yes. Copilot has official plugins for IntelliJ IDEA, PyCharm, WebStorm, GoLand and other JetBrains IDEs, with agent mode available. It also supports VS Code, Visual Studio, Eclipse, Xcode and Neovim.

What is Copilot’s IP indemnity?

On Business and Enterprise, GitHub provides intellectual property indemnity for Copilot-generated code, provided the public code filter is enabled. Individual plans do not include it.

Can I cancel Copilot anytime?

Yes. Monthly individual plans can be cancelled anytime and keep access until the end of the billing period. Business and Enterprise terms depend on your agreement.

Make an AI coding tool? Developers read this page when they're about to pay for one. Sponsor it →

DR

Written by the DevTools Review team

We research AI coding tools from vendor documentation, pricing pages, changelogs and public user reports, and re-check prices against vendors' own pages. We don't claim hands-on testing we haven't done. How we work.

Change tracker

Get told when a price changes

We re-check what all 12 tools cost against each vendor's own pricing page. When something moves — a price, a free tier, a limit — you get an email. When nothing moves, you get nothing.

Confirmation required. Unsubscribe in one click, from any email.