Copilot, Cursor or Claude Code? What Each One Costs Now
Copilot moved to credits, Cursor now belongs to SpaceX, and Claude Code comes with a chat plan. Here's what each one costs and who should pick which.
If you write code most days and want one AI assistant, start with GitHub Copilot Pro at $10 a month. It's the cheapest paid plan that doesn't fight you, and completions don't eat into your allowance. If you hand an AI whole tasks ("add this feature, run the tests, fix what breaks"), Claude Code on the $20 Pro plan is the better tool. Cursor is still a good editor, but since the SpaceX takeover its plans have changed so much that you should read the fine print before you pay.
That's the short version. The rest of this piece explains why, because all three changed how they charge this year.
What changed in 2026
Three things happened, and they all matter for your bill.
Copilot switched to credits. On June 1, 2026, GitHub retired "premium requests" on its individual plans and replaced them with GitHub AI Credits, which are billed by how many tokens you use. Each paid plan includes a credit amount roughly equal to its price, plus a "flex allotment" on top. Inline completions and next-edit suggestions are still free on paid plans. Chat and agent work is what spends credits.
Cursor was bought by SpaceX. SpaceX announced in June that it would buy Anysphere, the company behind Cursor, in an all-stock deal valued at $60 billion. The deal closed on August 14. Cursor's plans now sit alongside Grok, xAI's chatbot, and the cheapest paid tier only gives you Grok models.
Claude Code is part of Claude's normal plans. You don't buy it separately. Anthropic's Pro plan ($20 a month, or $17 a month billed yearly) includes it, and so do both Max tiers.
The prices side by side
Prices as of October 2026, from each company's pricing page:
| Entry paid plan | Mid plan | Top individual plan | Free tier | |
|---|---|---|---|---|
| GitHub Copilot | Pro, $10/mo | Pro+, $39/mo | Max, $100/mo | 2,000 completions + 50 chat requests a month |
| Cursor | Plus, $8/mo (Grok models only) | Premium, $30/mo | Ultra, $200/mo | Limited agent and chat |
| Claude Code | Pro, $20/mo | Max, from $100/mo | Max 20x | None (Free plan doesn't include it) |
The table hides the most important detail, though. These plans don't sell you a fixed number of answers any more. They sell you a pool of usage, and how fast you burn through it depends on which model you pick and how long your requests are. A plan that looks generous on paper can run dry in a week if you let an agent loose on a large codebase.
GitHub Copilot: the safe default
Who should pick it: most developers, especially if you live in VS Code or JetBrains and mainly want good autocomplete plus a chat panel.
Copilot's biggest strength is boring. It's everywhere: VS Code, Visual Studio, JetBrains IDEs, Neovim, the GitHub website and the command line. Your team probably already allows it. And since completions don't touch your credits on paid plans, the part you use all day costs nothing extra.
The Pro plan includes $10 of credits plus a $5 flex allotment. Pro+ ($39) bumps that to $39 plus $31 of flex, and it lets you hand tasks to third-party agents such as Claude and OpenAI Codex (in preview). The new Max plan ($100) is for people running agents for hours a day.
Pros
- Cheapest useful paid plan
- Unlimited completions on every paid tier
- Works in almost every editor
- Easy to get approved at work
Cons
- Credit billing makes heavy chat use hard to predict
- Agent mode is good, not great, for big multi-file changes
- GitHub paused new sign-ups for individual plans earlier this year; check the plans page before you count on it
Claude Code: best for handing off real work
Who should pick it: developers who want to describe a change in plain English and have it done across many files, including running tests and fixing what breaks.
Claude Code is a different shape of tool. It runs in your terminal (there are IDE extensions too) and works more like a colleague you give tasks to than an autocomplete. You tell it what you want, and it reads the code, makes a plan, edits files and runs your commands.
That style suits refactors, migrations, writing tests for old code and fixing bugs where you know the symptom but not the cause. It's less useful if what you want is a faster way to finish the line you're typing.
The pricing is simple to start with: it comes with the Pro plan you might already pay for to chat with Claude. Heavy users move up to Max (from $100 a month) because Pro's limits arrive quickly once you run long sessions.
Pros
- Built for multi-step, multi-file tasks rather than line-by-line help
- Included with a plan many people already have
- Works with any editor, because it lives in the terminal
Cons
- No inline autocomplete, so many people run it alongside Copilot
- Pro usage limits are tight for all-day agent work
- Takes a little practice to write good task descriptions
Cursor: still a great editor, but read the new plans
Who should pick it: people who want an AI-first editor and don't mind the plans changing under them.
Cursor built its name as a VS Code fork with AI wired into everything: tab completion that predicts your next edit, an agent panel and a fast way to chat with your whole codebase. All of that still works.
What changed is the business around it. After SpaceX closed the acquisition, Cursor's pricing page started listing Grok products next to the editor. The new $8 Plus plan is limited to Grok models. To use models from other companies you need Premium ($30) or higher, and third-party model usage is billed separately at API list prices. Super ($100) and Ultra ($200) give you 15x and 50x the Plus usage.
None of this makes Cursor bad. It does make it harder to predict. If you liked Cursor because it let you choose any model at a flat price, that deal no longer exists in the same form.
Pros
- The most polished AI-first editor experience
- Excellent tab completion for edits, not just new lines
- Cloud agents and code review on Premium and up
Cons
- Cheapest plan locks you to Grok
- Third-party models cost extra on top of the plan
- Pricing and branding may keep changing as it merges into SpaceX's AI business
So which one?
- You want better autocomplete and a chat panel: Copilot Pro, $10.
- You want to give the AI whole tasks: Claude Code with Claude Pro, $20. Add Copilot Pro if you miss autocomplete; plenty of people pay for both.
- You want everything inside one AI editor and you're happy with Grok: Cursor Plus, $8. If you need other models, budget for Premium at $30 plus usage.
- You're a student or just curious: start with Copilot Free. 2,000 completions a month is enough to find out whether you like it.
One last thing. Whatever you pick, check your usage page after the first week. Every one of these tools now bills by consumption in some form, and the first week tells you more about your real cost than any pricing table, including this one.
FAQ
Is GitHub Copilot still $10 a month?
Yes. Copilot Pro is $10 a month. Since June 2026 it includes $10 of AI credits plus a flex allotment, and code completions don't use credits.
Do I need a separate subscription for Claude Code?
No. It's included in Claude Pro ($20 a month, or $17 a month billed yearly) and in the Max plans. The Free plan doesn't include it.
Did SpaceX really buy Cursor?
Yes. SpaceX agreed to buy Anysphere, Cursor's parent company, in an all-stock deal valued at $60 billion. It closed on August 14, 2026.
Can I use Claude or GPT models in Cursor?
Yes, but not on the $8 Plus plan, which is Grok-only. On Premium and higher, third-party models are available and billed at API list prices on top of your plan.
Related guides
- What Is an AI Agent? A Plain-English Explanation
- How to Learn Programming With AI Without Fooling Yourself
- How to Run an AI Model on Your Own Laptop (Free, Private, Offline)
