GitHub Copilot Pro+ costs $39 a month, and for most of 2026 the question was simple: is it worth $29 more than Pro? That question changed on June 1. GitHub stopped counting premium requests and started charging AI credits for chat and agent work, then added a $100 Max plan above Pro+. So the real choice now is about two things: how many credits you burn, and which models you need. This review walks through every GitHub Copilot plan as of October 2026, what the credits actually buy, which models sit behind the Pro+ paywall, and who should pay for which tier.
⚡ TL;DR: The Bottom Line
What It Is: GitHub’s AI coding assistant: unlimited code completions on paid plans, plus chat, agent mode, a cloud agent, code review and the Copilot CLI, all paid for with monthly AI credits.
Price: Free / Pro $10 / Pro+ $39 / Max $100 a month. Teams: Business $19 and Enterprise $39 per seat.
What Pro+ Adds: $70 of monthly credits instead of $15, and the premium models Pro doesn’t get: Claude Opus 5.5, Claude Fable 5.1, GPT-6 Sol and GPT-6.1 Sol.
Our Take: Pro at $10 costs half of the $20 coding plans from Anthropic and OpenAI. New Cursor sign-ups now pay $30 a month for Premium, or $8 for Plus with Grok models only. Pro+ is worth it once you either need the top Claude or GPT-6 models, or your extra usage on Pro would pass $29 a month (about $44 of credits in total).
⚠️ The Catch: Part of every plan’s credits is a “flex allotment”, and GitHub says flex allotments may change over time. On Pro+ that flex part is $31 of the $70.
📑 Quick Navigation
The Bottom Line: 60-Second Verdict
Three quick recommendations, depending on where you are:
- New to Copilot? Start on Free, then move to Pro at $10 a month. Pro includes $15 of AI credits, unlimited completions, the cloud agent and code review, and models like Claude Sonnet 5.5, GPT-6 Luna and Grok 4.7.
- Thinking about Pro+ ($39 a month)? Pay the extra $29 only if you need the premium models (Claude Opus 5.5, Claude Fable 5.1, GPT-6 Sol, GPT-6.1 Sol) or your extra credit spending on Pro would pass $29 a month. Pro+ comes with $70 of credits, $55 more than Pro.
- Running agents all day? Look at Max at $100 a month. It includes $200 of credits and “priority access to premium models”. Compare it with paying for extra credits on Pro+ first: credits cost $0.01 each either way.
The rest of this review covers what changed since April, what every plan includes, how the credits work, and how Copilot compares with Cursor, Claude Code and OpenAI Codex at the $20 mark.
What Changed Since April 2026
Three changes since this review’s April update decide any buying choice today.
1. Sign-ups are open again
In April, GitHub paused new sign-ups for its individual plans ahead of a billing change. GitHub’s June 16 update said it was reopening sign-ups for Student, Pro, Pro+ and Max “gradually over the next couple of weeks”. As of October 4, 2026, GitHub’s plans page shows a Get started button on Free, Pro, Pro+ and Max.
2. Premium requests became AI credits on June 1
GitHub announced on April 27 that premium request units would be replaced by GitHub AI Credits from June 1, 2026, and that has been the system on every monthly plan since. Annual Pro and Pro+ subscribers keep premium request pricing until their plan expires. One credit is worth $0.01. Chat, agents, the Copilot CLI, Spaces and Spark spend credits. Code completions and next edit suggestions don’t, and stay unlimited on every paid plan.
3. A new Max plan, and a new model list
Copilot Max costs $100 a month and sits above Pro+. GitHub describes it as built for “sustained agent-driven workflows”. The model picker also turned over: the April lineup (Opus 4.7, GPT-5.5 as the top OpenAI pick) is now led by Claude Opus 5.5, Claude Fable 5.1 and the GPT-6 family. Those top models are Pro+ and up. GitHub’s supported models page warns that “model availability is subject to change”.
What GitHub Copilot Actually Does in 2026
GitHub Copilot started as autocomplete. Today it is a set of AI tools that live where your code already lives: your editor, your terminal and GitHub itself.
The Copilot surfaces
- Code completions: suggestions as you type, plus next edit suggestions. Unlimited on paid plans and free of credits. Copilot Free caps them at 2,000 a month.
- Chat: ask questions about your code inside the editor, on GitHub.com or on mobile. Spends credits, more for bigger models.
- Agent mode: describe a change and Copilot plans it, edits files and runs commands in VS Code, Visual Studio, JetBrains, Eclipse or Xcode. Available on every plan, including Free.
- Cloud agent and code review: assign an issue to Copilot and it works on its own and opens a pull request; it also reviews pull requests. Both start at Student and Pro (Free only gets “Review selection” in VS Code).
- Copilot CLI and Spark: Copilot in your terminal on every plan, and Spark, a plain-English app builder in public preview, listed for Pro+ and Enterprise.
The structural advantage over rivals is still editor breadth. Copilot works in VS Code, Visual Studio, the JetBrains IDEs, Vim and Neovim, Xcode and Eclipse, plus GitHub.com and the CLI, with completions in all of them. Cursor’s Tab completions live in its own editor; its agent reaches JetBrains, Neovim and Zed through the Agent Client Protocol on paid plans.
What Your First Five Minutes Look Like
Based on GitHub’s plan docs, here is what a new subscriber sees on day one.
- Minute 1: pick a plan on the plans page. Free needs no payment. Verified students get Copilot Student free, and verified teachers and maintainers of popular open source projects may get Pro free.
- Minute 2: install the Copilot extension in your editor and sign in with your GitHub account. Completions start right away and never touch your credits on a paid plan.
- Minute 3: open chat and look at the model picker. On Free and Student it only offers auto model selection. Pro shows “a selection of models”; Pro+ and Max add the premium ones.
- Minute 4: try agent mode on a small change. This is where credits start to move: a quick question to a light model costs a fraction of a credit, a long agent session on a frontier model costs more.
- Minute 5: open your Copilot settings and set a budget. GitHub shows your usage and reset date there, with alerts at 75%, 90% and 100% of any budget you set.
Every GitHub Copilot Plan, Explained (October 2026)
There are seven Copilot plans: five for individuals and two for organizations. Here is the full lineup from GitHub’s plan docs, checked October 4, 2026.
- Copilot Free ($0): limited chat and agent usage, auto model selection only, 2,000 completions a month. No cloud agent.
- Copilot Student (free for verified students): an allowance of AI credits, auto model selection only, cloud agent and code review included, third-party agents excluded.
- Copilot Pro ($10 a month): $15 of AI credits a month ($10 base plus $5 flex), unlimited completions, “a selection of models”.
- Copilot Pro+ (Pro Plus, $39 a month): $70 of credits ($39 base plus $31 flex), access to premium models, Spark (preview) and audit logs.
- Copilot Max ($100 a month): $200 of credits ($100 base plus $100 flex) and priority access to premium models.
- Copilot Business ($19 per seat a month): 1,900 credits ($19) per user, organization policies, content exclusion and audit logs.
- Copilot Enterprise ($39 per seat a month): 3,900 credits ($39) per user and priority access to premium models, for companies on GitHub Enterprise Cloud.
💡 Key Takeaway: The price you see is the entry fee, and the credits are what you actually spend. Pro turns $10 into $15 of credits, Pro+ turns $39 into $70, and Max turns $100 into $200. If you stay under your allowance, your bill is the sticker price.
Features and Models: What Each Plan Unlocks
Agent mode (every plan)
Agent mode is included on every plan, Free too. You describe a feature in plain language and Copilot plans the steps, edits several files, runs terminal commands and reports back. It works with MCP servers and custom instructions on every plan. The difference between plans isn’t access to agent mode. It is how many credits you have to spend in it, and which models it can use.
The model picker: the real Pro vs Pro+ line
This is the biggest reason to pick Pro+. According to GitHub’s plan docs, Pro includes Claude Haiku 4.5, Claude Sonnet 5 and Sonnet 5.5, Gemini 3.7 Flash and 3.8 Flash, GPT-5 mini, GPT-5.3-Codex, GPT-5.4, GPT-5.4 mini, GPT-5.6 Luna and Terra, GPT-6 Luna, Grok 4.5 to 4.7, Kimi K3 and MAI-Code-1.1-Flash.
Pro+ and Max add the premium tier: Claude Opus 4.8, Opus 5 and Opus 5.5, Claude Fable 5 and 5.1, GPT-5.5, GPT-5.6 Sol, GPT-6 Astra, GPT-6 Sol and GPT-6.1 Sol. So if your work needs Anthropic’s Opus or Fable models, or OpenAI’s GPT-6 Sol, Pro won’t do.
Copilot cloud agent and code review (Pro and up)
Assign work to Copilot and it researches, plans and writes code on its own, with or without a pull request. Code review checks pull requests on GitHub and file diffs in your editor. Both are in Student, Pro, Pro+ and Max, and both spend credits.
GitHub Spark (preview)
Spark builds small web apps from a plain-English description. It is in public preview, and GitHub’s plan docs list it for Pro+ and Enterprise. Good for prototypes and internal tools, not a replacement for a real codebase.
Editor breadth (every plan)
A single Copilot subscription works in VS Code, Visual Studio, JetBrains, Vim and Neovim, Xcode and Eclipse, plus GitHub.com, GitHub Mobile and the CLI. For teams with mixed editors, or anyone who isn’t on VS Code, this matters more than any single feature.
🔍 REALITY CHECK
Marketing Claims: “Pro+ gives you access to premium models” (GitHub’s plan description for the $39 tier).
What the Fine Print Says: True, and the list is long: Opus 5.5, Fable 5.1, GPT-6 Sol and more are Pro+ only. But premium models also spend credits faster, and choosing a larger context window or higher reasoning level uses more credits again. The $70 allowance can go quickly if you use the biggest models for every task.
Verdict: Pro+ is the right plan if you need those models for hard problems. Use the lighter models for routine work so your credits last the month.
Pricing Breakdown: Pro vs Pro+ vs Max (October 2026)
| Plan | Monthly | AI Credits / Month | Models | Best For |
|---|---|---|---|---|
| Free | $0 | Limited | Auto selection only | Trying Copilot |
| ProVALUE | $10 | $15 (1,500) | A selection (Sonnet 5.5, GPT-6 Luna, Grok 4.7) | Most individual developers |
| Pro+ | $39 | $70 (7,000) | Premium (Opus 5.5, Fable 5.1, GPT-6 Sol) | Daily agent users, top models |
| Max | $100 | $200 (20,000) | Priority access to premium | All-day agent workflows |
| Business | $19/seat | $19 (1,900) | Premium | Teams that need admin controls |
| Enterprise | $39/seat | $39 (3,900) | Priority access to premium | GitHub Enterprise Cloud orgs |
Here’s the maths that matters. Pro+ costs $29 a month more than Pro and includes $55 more in credits ($70 vs $15). Max costs $61 more than Pro+ and includes $130 more ($200 vs $70). Measured as credits per dollar of subscription, Pro gives $1.50, Pro+ about $1.79 and Max $2.00. The bigger plans are better value per dollar, but only if you actually use the credits.
One caution on those numbers: part of each allowance is the flex allotment ($5 on Pro, $31 on Pro+, $100 on Max), which GitHub calls “variable additional usage on top of your base” and says may change over time. The base part matches the price you pay.
📊 Price vs Included AI Credits (October 2026)
Monthly price and the value of the AI credits each individual plan includes, in USD. Source: GitHub’s plans page, October 4, 2026.
📬 Find this review useful?
Get our weekly developer-tool updates: pricing changes, new plans and model changes.
Copilot vs Cursor vs Claude Code vs Codex (Head-to-Head)
| Specification | GitHub Copilot Pro | Cursor (Pro existing / Premium new) | Claude Code | OpenAI Codex |
|---|---|---|---|---|
| Monthly price | $10 | $20 for existing subscribers; new sign-ups: Premium $30 (Plus $8 runs Grok models only) | $20 (in Claude Pro) | $20 (in ChatGPT Plus) |
| How usage is counted | $15 of AI credits, then $0.01 per credit | “Extended limits on Agent” | Claude Pro usage limits | ChatGPT Plus usage limits |
| Where it runs | 6 editor families + GitHub.com + CLI | Cursor’s editor, CLI, cloud agents; agent in JetBrains via ACP | Included in Claude Pro and Team | Web, CLI, IDE extension, iOS |
| Models | Anthropic, OpenAI, Google, xAI and more | “Access to frontier models”; new plans: Grok included, others at API list price | Anthropic only | OpenAI only |
| Team plan | Business $19/seat | Teams $40/seat for existing teams; new sign-ups from $30/seat | Claude Team | ChatGPT Business |
Reading the table: the cheapest paid GitHub Copilot subscription, Pro at $10, is half the price of Claude Pro or ChatGPT Plus ($20 each), and half of what existing Cursor Pro subscribers pay. Cursor’s Pro plan has been closed to new sign-ups since October 12: new customers get The X Subscription, where Premium costs $30 a month and Plus costs $8 but runs only Grok models in Cursor. On Premium and up, Claude and GPT models in Cursor are billed separately at API list prices. Copilot also works in more editors than the others and has GitHub’s cloud agent and code review built in. (ChatGPT Free and Go include limited Codex too, and Cursor has a Free plan with limited agent requests.) Claude Code and Codex come bundled with something else you may already pay for: Claude Code with Claude Pro, Codex with ChatGPT Plus. If you already pay for one of those subscriptions, try its coding agent before adding a second bill.
What we have dropped from this table since April: the speed and SWE-bench figures. They came from third-party tests of older models, and every tool in the table has changed models since. We’ll add benchmark rows back only when a vendor or a well-documented public benchmark publishes them for the current models.
Who Should Pay for What (October 2026 Decision Guide)
- You code a few hours a week: Copilot Free or Pro. Pro’s unlimited completions plus $15 of credits covers light chat and the occasional agent task.
- You live in agent mode: Pro+ at $39. You get almost five times Pro’s credits ($70 vs $15) and the premium models.
- You need Opus 5.5, Fable 5.1 or GPT-6 Sol inside your editor: Pro+ is the cheapest individual Copilot plan that includes them (Business seats include them too, but with $19 of credits).
- You run agents all day, every day: Max at $100, or Pro+ with a paid-usage budget. Both charge $0.01 per extra credit, so Max is the better deal only if you would spend more than $61 a month of extra credits on Pro+.
- You are a team of two or more: Copilot Business at $19 per seat for admin controls, content exclusion and audit logs. Note that Business includes $19 of credits per user, much less than Pro+.
- You already pay for Claude Pro or ChatGPT Plus: try Claude Code or Codex first. You may not need a second coding subscription.
- You use an editor other than VS Code: Copilot is the safest pick. It covers JetBrains, Visual Studio, Vim, Neovim, Xcode and Eclipse.
🔍 REALITY CHECK
Marketing Claims: “GitHub Copilot Pro is just $10 a month” (GitHub’s entry-tier pricing).
What the Fine Print Says: The $10 is real, and it includes $15 of AI credits. Completions are unlimited and free. But once chat and agent work use up the allowance, you either wait for the monthly reset, switch to a cheaper model, or set a budget and pay $0.01 per credit (a $10 budget buys 1,000 credits).
Verdict: For light users, $10 is the whole bill. Heavy agent users should set a budget in Copilot settings so the bill can’t surprise them.
How GitHub AI Credits Work (Since June 1, 2026)
The June 1 switch was the biggest change to Copilot’s pricing in years. Here is how the system works now, straight from GitHub’s plans page.
What a credit is: one GitHub AI credit equals $0.01. Each plan includes a monthly allowance that resets every month.
What spends credits: chat, agent mode, the cloud agent, code review, the Copilot CLI, Spaces and Spark. Usage is calculated from tokens, so the model you pick and the size of the task decide the cost. Larger context windows and higher reasoning levels use more.
What doesn’t: code completions and next edit suggestions. They stay unlimited on every paid plan.
When you run out: wait for the next cycle, switch to a lighter model, or set a dollar budget for extra usage and keep going. On Business and Enterprise, admins decide whether extra paid usage is allowed. If it isn’t, Copilot pauses until the next cycle.
Who it costs more: the old system counted requests. The new one counts tokens, so a long agent session on a frontier model is billed for everything it reads and writes. GitHub hasn’t published a typical cost per task, so the safest move is to watch your usage page for the first month and set a budget.
FAQs
Can I sign up for GitHub Copilot Pro right now?
Yes. GitHub paused new individual sign-ups in April 2026 and began reopening Student, Pro, Pro+ and Max gradually from June 16. As of October 4, 2026, the plans page shows a Get started button on Free, Pro, Pro+ and Max.
How much does GitHub Copilot cost in 2026?
Free costs $0, Pro $10 a month, Pro+ $39 and Max $100. For organizations, Business is $19 and Enterprise $39 per seat per month. Students verified by GitHub get Copilot Student free.
Is GitHub Copilot Pro+ worth $39 a month over Pro at $10?
It is if you need the premium models (Claude Opus 5.5, Claude Fable 5.1, GPT-6 Sol, GPT-6.1 Sol) or your extra credit spending on Pro would pass $29 a month. Pro+ includes $70 of credits a month. If neither applies, Pro is the better buy.
What are GitHub AI Credits?
They are how Copilot charges for AI usage since June 1, 2026. One credit is worth $0.01. Chat, agents, code review and the CLI spend credits based on tokens; completions and next edit suggestions don’t. Pro includes 1,500 credits ($15), Pro+ 7,000 ($70) and Max 20,000 ($200).
What is GitHub Copilot Max?
Max is the top individual plan at $100 a month. It includes $200 of AI credits and priority access to premium models, and GitHub says it is built for sustained agent-driven workflows.
Which models can I use on Copilot Pro?
Pro includes Claude Haiku 4.5, Sonnet 5 and Sonnet 5.5, Gemini 3.7 Flash and 3.8 Flash, several GPT-5.x models, GPT-6 Luna, Grok 4.5 to 4.7, Kimi K3 and MAI-Code-1.1-Flash. Opus, Fable, GPT-6 Sol and GPT-6 Astra need Pro+ or higher.
Does GitHub Copilot train on my code?
GitHub states it does not use Copilot Business or Enterprise data to train its models. If that guarantee matters for your work, use an organization plan.
Can I use GitHub Copilot in editors other than VS Code?
Yes. Copilot works in Visual Studio, the JetBrains IDEs, Vim and Neovim, Xcode and Eclipse, plus GitHub.com, GitHub Mobile and the Copilot CLI.
✅ What We Liked
- ✓ Low entry price: $10 a month with $15 of credits
- ✓ Unlimited completions on every paid plan, outside the credit meter
- ✓ Models from Anthropic, OpenAI, Google and xAI in one subscription
- ✓ Works in six editor families plus GitHub.com and the CLI
- ✓ Native GitHub workflow: cloud agent, pull requests, code review
❌ What Fell Short
- ✗ Top models (Opus 5.5, Fable 5.1, GPT-6 Sol) are Pro+ and up
- ✗ Agent work on big models can drain credits fast
- ✗ Part of each allowance is “flex” that GitHub says may change
- ✗ Business seats include far fewer credits than Pro+
A low-cost entry into AI coding, with unmatched editor breadth and native GitHub integration. One star off because the best models sit behind Pro+ and heavy agent use now depends on a credit meter with a changeable flex part.
💡 Key Takeaway: Before you upgrade, open your Copilot settings and look at last month’s credit usage. If your extra usage on Pro would stay under $29 a month (about $44 of credits in total), stay on Pro. Above that, or if you need Opus 5.5 or GPT-6 Sol, Pro+ is the better deal. Only consider Max if your extra spend on Pro+ would pass $61 a month.
The Final Verdict
GitHub Copilot is still the easiest AI coding tool to recommend in 2026, and since sign-ups reopened, new subscribers can sign up again. Pro at $10 costs half of the $20 plans from Anthropic and OpenAI, and new Cursor sign-ups now pay $30 a month for Premium, or $8 for Plus with Grok models only. For $10 you get unlimited completions, $15 of AI credits, the cloud agent and code review, in almost any editor.
GitHub Copilot Pro+ at $39 earns its price for two kinds of people: those who need the premium models (Claude Opus 5.5, Claude Fable 5.1, GPT-6 Sol) and those who use agent mode every day and outgrow Pro’s credits. Max at $100 is for the few who run agents all day. Everyone else should start on Pro, watch the usage page for a month, and let the numbers decide.
Related Reading
- Claude Code Review: Anthropic’s terminal-first coding agent, included with Claude Pro
- Cursor AI Review 2026: the AI-first editor, now sold through The X Subscription
- Claude Code vs Cursor 2026: the two main alternatives to Copilot, side by side
- Google Antigravity Review: Google’s agentic IDE
- ChatGPT Codex Review: OpenAI’s coding agent, included with ChatGPT Plus
- Lovable Review: the no-code app builder that competes with GitHub Spark
- Gemini CLI Extensions: Google’s terminal AI
- The Complete AI Tools Guide: a buyer’s guide to picking the right AI tool
Founder of AI Tool Analysis. Reads every pricing page, help article and footnote so you don’t have to. Covering AI tools since 2025. See how we research →
Stay Updated on AI Coding Tools
The AI coding category is shifting fast. Subscribe for plain-English updates on new releases, plan changes and pricing, delivered every Thursday at 9 AM EST.
- ✅ Sourced Facts: Every price and limit checked against the vendor’s own pages
- ✅ Price Tracking: Know when tools drop prices or restructure tiers
- ✅ Plan Changes: First to know when tiers open, close or change
- ✅ Comparison Updates: Side-by-side plan comparisons as the market shifts
- ✅ No Hype: Just the AI news that actually matters for your work
Free, unsubscribe anytime.
Last Updated: October 13, 2026 | Facts checked: October 13, 2026 (plans, prices, credits and model access against GitHub’s plans page and Copilot docs) | Next Review Update: November 12, 2026, or on the next plan change.
Have a tool you want us to review? Suggest it here | Questions? Contact us