Cursor vs GitHub Copilot vs Claude Code: 2026 Pricing

Side-by-side pricing tiers for Cursor, GitHub Copilot and Claude Code, grouped by monthly price and usage allowance

Cursor, GitHub Copilot and Claude Code all changed how they charge within the past five months, and the three now meter usage in three different units. This guide answers a practical question: what does each price point actually buy, and where does the monthly fee stop covering your work?

This comparison is based on official documentation, published specifications, and vendor pricing verified as of October 26, 2026. Every vendor page cited below was opened directly. Figures marked "derived estimate" are my own arithmetic on published rates, not vendor numbers.

Three plans, three billing units

Dollar prices are not comparable until you know what the dollars meter. The table below normalizes the structure rather than the price.

	Cursor	GitHub Copilot	Claude Code
What the plan meters	One shared pool, expressed as multiples of the Plus plan (Premium 4x, Super 15x, Ultra 50x)	AI credits worth $0.01 each, consumed by tokens at each model's API rates	A rolling five-hour session limit plus a weekly limit, shared with claude.ai
Reset	Weekly	Monthly	Five-hour and weekly windows
Models inside the plan	Cursor's own models only (Grok 4.7, 4.6, 4.5, Composer 2.5)	Model list spans Anthropic, OpenAI, Google, xAI, Moonshot and Microsoft; availability varies by plan and feature	Claude models only
Other vendors' models	Premium and above: available at API list price from prepaid credit; not included	Drawn from the same credits	Not available
When you run out	Wait for weekly reset, or spend prepaid credit at per-model rates	Buy extra credits; no automatic fallback to a cheaper model	Wait, upgrade, or buy usage credits at standard API rates

Sources: Cursor's Models & Pricing docs, GitHub's usage-based billing announcement, and Anthropic's Claude Code plan article. The practical consequence: only Copilot publishes its allowance in dollars, so it is the only one where you can compute what the fee buys.

The price ladder, tier by tier

Tier	Cursor	GitHub Copilot	Claude
Entry	Plus $8/mo ($80/yr): Grok models only	Pro $10/mo: $15 in credits	Pro $20/mo ($17 on annual billing)
Mid	Premium $30/mo: 4x Plus	Pro+ $39/mo: $70 in credits	Max, from $100/mo: 5x or 20x Pro usage
High	Super $100/mo: 15x Plus	Max $100/mo: $200 in credits	Same Max plan; a separate price for the 20x level is not shown on the pricing page
Top	Ultra $200/mo: 50x Plus	None	None
Teams	$30, $100 or $200 per seat	Business $19/seat, Enterprise $39/seat	Standard $20 annual / $25 monthly; Premium $100 / $125

Prices come from Cursor's pricing page, GitHub's plans page and Claude's pricing page. Cursor's pricing page lists only monthly prices, while its help article gives annual prices of $80, $300, $1,000 and $2,000 for the four paid individual plans. Claude Code is included in Pro, Max, Team and Enterprise, and not in Free.

What the included credit buys (derived estimate)

Copilot's credits can be converted into a workload. Take one agent step of 100K input and 20K output tokens. At Anthropic's published rates, that costs $0.80 on Claude Opus 5.5 ($4 and $20 per million tokens) or $0.40 on Sonnet 5.5 ($2 and $10). GitHub says credits are priced at each model's published API rates, so the arithmetic is a fair first approximation.

Copilot plan	Included credit	Opus 5.5 steps	Sonnet 5.5 steps
Pro	$15	18	37
Pro+	$70	87	175
Max	$200	250	500
Business (per seat, pooled)	$19	23	47
Enterprise (per seat, pooled)	$39	48	97

Two caveats. Copilot code review also consumes GitHub Actions minutes, and Copilot's own multipliers may differ from list rates, so treat these as upper bounds. On Cursor, the same Opus step is not included at any tier: on Premium and above it bills at API list price from prepaid credit, so $0.80 per step from the first request. Claude Code publishes no absolute allowance, only multiples of Pro, so no step count can be derived for it.

Cursor's restructure: who pays more

Cursor's help article maps legacy plans to the new ones. Applying the published prices gives these changes (derived estimate):

Pro $20 to Premium $30: +50%.
Pro+ $60 to Super $100: +67%.
Teams Standard $40 to the $30 seat: -25%. Teams Premium $120 to the $100 seat: -17%.

Switching is optional: Cursor states that legacy plans continue until you change or cancel, but moving to the new subscription cancels your Cursor plan with a prorated refund and is permanent. The structure changed too: legacy plans used two monthly pools plus on-demand usage, and the new ones use one weekly pool shared with Grok. The help page gives no calendar date for the change.

Where each plan stops working for you

Cursor Plus: only Grok models are available. If your workflow depends on Claude or GPT models, Plus does not cover it.
Cursor, all paid plans: subscriptions are sold only directly through X Subscription Corp., and the pricing page warns that reseller purchases are unauthorized. Teams and Enterprise add a $0.25 per million token surcharge on eligible third-party models.
Copilot: GitHub removed fallback to a cheaper model when you hit limits. Delegating tasks to third-party agents such as Claude or OpenAI Codex is limited to Pro+ and Max, and is in preview. Per GitHub's announcement, annual Pro and Pro+ subscribers stay on the older premium-request pricing until their plan expires.
Claude Code: limits are shared with claude.ai, including use inside IDEs such as Cursor. If an ANTHROPIC_API_KEY variable is set, Claude Code bills API charges instead of drawing on your subscription. Monthly API credits that come with Max cannot be used in Claude Code, and the Max plan has a separate weekly limit for Fable.

Where the published numbers disagree

Copilot credits. GitHub's announcement post lists included credits equal to the plan price (Pro $10, Pro+ $39). The current plans page adds a flex allotment on top: Pro $15, Pro+ $70, Max $200. Use the current page.
Cursor annual pricing. The pricing page shows monthly prices only, while the help article lists annual prices. Both are Cursor's own pages.
Copilot Free models. GitHub's FAQ says Free users get entry-level models, but the plans page lists the same models for all individual plans. I could not extract per-plan model checkmarks, so this guide does not claim any.
Claude Max pricing. The page says "from $100" with a choice of 5x or 20x Pro usage, but does not give a separate 20x price in the text I could read.

Which plan for which situation

You want the most Claude and GPT usage per dollar, in one bill: Copilot Pro+ or Max. They are the only plans here where one dollar-denominated allowance covers several vendors' models.
You live in the terminal with Claude models: Claude Pro or Max. The fee covers Claude Code and claude.ai from the same limits, which is efficient if you accept that heavy chat use reduces coding headroom.
You want Cursor's editor and Grok models: Cursor Premium. Budget separately for any Claude or GPT usage, because none is included.
You need cheap, light assistance: Copilot Pro at $10 has unlimited inline completions, which do not consume credits.

FAQ

Is Claude Code included in Claude Pro?

Yes. Anthropic's pricing page lists Claude Code in Pro, Max, Team and Enterprise, with usage shared with the rest of Claude. Free does not include it.

Does Cursor Premium include Claude models?

No. Cursor's pricing page says third-party models are not included in the monthly price. On Premium and above they are billed at API list price from prepaid credit.

Do Copilot code completions use credits?

No. GitHub states that code completions and next edit suggestions do not consume AI credits. Agent mode, code review and the CLI do.

Last verified: October 26, 2026. All three vendors changed pricing structures this year, so confirm current rates on their pages before subscribing.