Cursor vs GitHub Copilot vs Claude Code vs Codex: Prices Compared

The cheapest paid way into an AI coding agent in September 2026 is GitHub Copilot Pro at $10 a month, while Cursor Pro, Claude Pro (which includes Claude Code) and ChatGPT Plus (which includes Codex) all start at $20. The bigger differences are in how each one meters usage: Copilot uses dollar-denominated AI credits, Cursor splits usage into two pools, and Claude Code and Codex share a rolling allowance with the rest of their chat subscriptions.

Cursor vs Copilot vs Claude Code vs Codex at a glance

These four tools overlap heavily, but they come from different directions. Cursor is an AI-first code editor with its own agent and cloud agents. GitHub Copilot lives inside existing IDEs, GitHub.com and a CLI. Claude Code is Anthropic’s agent that started in the terminal and now also runs in IDEs, a desktop app, the web and Slack. Codex is OpenAI’s coding agent, bundled with ChatGPT plans and available on the web, in a CLI, in an IDE extension and on iOS.

Cursor pricing and plans

According to Cursor’s pricing page, the individual plans are Hobby (free, no credit card, limited Agent requests), Pro at $20 a month, Pro+ at $60 and Ultra at $200. Teams costs $40 per user per month, and Enterprise is custom-priced. Cursor itself recommends Pro+ for daily agent users and Ultra for agent power users.

The structure is unusual. Cursor’s models and pricing documentation explains that Pro, Pro Plus and Ultra each include two separate usage pools that reset with your billing cycle:

• Cursor Models: “significantly more included usage” for Grok 4.7, Grok 4.6, Grok 4.5 and Cursor’s own Composer 2.5.
• Other Models: third-party models such as Claude Opus 5.5, Claude Sonnet 5, Gemini 3.8 Flash, Muse Spark 1.3 and OpenAI’s GPT-5.6 family, charged against your allowance at each model’s API price, with optional paid overage.

The Grok emphasis follows ownership: SpaceX completed its $60 billion all-stock purchase of Cursor’s parent Anysphere on August 14, 2026, and Cursor now sits in the SpaceXAI division. If you mostly use SpaceXAI’s models, Cursor’s included usage goes considerably further; see our Grok 4.7 pricing and benchmarks explainer for what the newest one offers. If you want to live in Claude or GPT models, you are drawing on the smaller pool at API rates.

All paid individual plans include unlimited Tab completions, extended agent limits, Bugbot code review and Cloud Agents. In India, Cursor also sells a Start plan at ₹649 a month, tax inclusive, which covers the Cursor Models pool only, runs Grok models at a fixed medium effort level and excludes Bugbot, Auto and on-demand usage.

GitHub Copilot pricing and plans

GitHub has moved Copilot to a credit system. The official Plans for GitHub Copilot documentation lists these tiers:

GitHub’s billing documentation for individuals spells out how credits work. One AI credit equals $0.01, and every interaction that uses a model is priced by input, output and cached tokens for that model. A quick chat question with a light model might cost a fraction of a credit, while a long cloud-agent session with a frontier model costs more. Base credits are spent first, then the flex allotment, and unused credits do not roll over; the allowance resets at midnight UTC on the first of each month.

Two details work in Copilot’s favour. Code completions and next edit suggestions are not billed in credits and stay unlimited on paid plans. And Copilot’s model menu is among the widest of the four: GitHub’s plan page lists models from Anthropic (including Claude Opus 5.5), OpenAI (including GPT-6 Sol), Google (Gemini 3.8 Flash), SpaceXAI (Grok 4.7) and Moonshot (Kimi K3). It also offers third-party agents, including OpenAI Codex and Anthropic Claude, in public preview. Verified students get Copilot Student for free, and verified teachers and maintainers of popular open-source projects may qualify for free Copilot Pro.

Claude Code pricing and plans

Claude Code is not sold on its own for individuals. It is included in paid Claude subscriptions, or you can use it with pay-as-you-go API billing through a Claude Console account. Anthropic’s Claude pricing page lists:

• Free: no Claude Code access.
• Pro: $20 billed monthly, or $17 a month with an annual subscription ($200 billed up front). Includes Claude Code.
• Max 5x: $100 a month, five times Pro’s usage.
• Max 20x: $200 a month, twenty times Pro’s usage.
• Team: standard seats at $20 per seat per month billed annually ($25 monthly); premium seats with five times the usage at $100 annually ($125 monthly).
• Enterprise: $20 per seat per month billed annually, plus usage at API rates.

The Claude Code product page highlights where it runs: the terminal, native extensions for VS Code (which Anthropic says also work in Cursor) and JetBrains IDEs, the Claude desktop app, the web and Slack. Anthropic says Claude Code runs locally in your terminal and talks directly to model APIs “without requiring a backend server or remote code index,” and it asks permission before changing files or running commands. It also now runs in an “auto mode” by default on Pro, Max and Team plans, which Anthropic says lets it work longer while still catching risky commands.

The trade-off is model choice: Claude Code is built around Anthropic’s own models, including the new Claude Opus 5.5, while Copilot and Cursor let you pick between several vendors.

OpenAI Codex pricing and plans

Codex is included in every ChatGPT plan, and OpenAI’s Codex pricing documentation notes that it shares usage with ChatGPT Work. The tiers:

• Free ($0) and Go ($8 a month): GPT-6 Luna at standard speed in the desktop app, subject to rollout. Go is aimed at “lightweight coding tasks.”
• Plus ($20 a month): Codex on the web, in the CLI, in the IDE extension and on iOS, plus cloud integrations such as automatic code review and Slack. Includes GPT-6 Sol and GPT-6 Luna.
• Pro (from $100 a month): a choice of 5x or 20x Plus’s Codex usage.
• Business ($20 per user per month, billed annually; $25 monthly): for two or more users, with admin controls and no training on business data by default.
• Enterprise and Edu: contact sales.
• API key: Codex in the CLI, SDK or IDE extension billed at API rates, without cloud features such as GitHub code review or Slack.

OpenAI publishes rough local-message estimates per five-hour window rather than hard caps. For GPT-6 Sol, it lists 15 to 150 messages on Plus, 70 to 700 on Pro 5x and 300 to 3,000 on Pro 20x; the lighter GPT-6 Luna ranges from 350 to 3,000 on Plus. Weekly limits may also apply. OpenAI also notes that GPT-5.5 retires from ChatGPT and Codex on all plans on October 14, 2026. Our GPT-6 Sol and Luna explainer covers the models themselves.

Which AI coding tool is cheapest?

It depends on what you count as “using” it.

• For autocomplete-heavy work: Copilot Pro at $10 is the cheapest paid plan, and completions do not consume credits. Cursor’s paid plans also include unlimited Tab completions, but start at $20.
• For occasional agent tasks: ChatGPT Go at $8 gets you Codex with GPT-6 Luna in the desktop app, though OpenAI positions it for lightweight work only.
• For daily agentic coding on a single subscription: the $20 tier is where all four become genuinely usable. Claude Pro is the cheapest if you pay annually, at $17 a month.
• For all-day power use: Copilot Max ($100, 20,000 credits) and ChatGPT Pro’s 5x option (from $100) are the cheapest heavy-use tiers on paper, while Cursor Ultra and Claude Max 20x both cost $200.

Because every vendor meters differently, the per-month sticker price is only a rough guide. A long agent run on a frontier model can burn through a $10 credit allowance in far fewer sessions than the same work on a lighter model.

Which is best for teams?

Copilot is the natural default for organisations already paying for GitHub, since seats, policies and billing sit in the same place. Cursor has the most expensive standard seat but bundles agentic code review and a team marketplace for internal rules, skills and plugins. Claude Team and ChatGPT Business make most sense when the team also wants the vendor’s chat assistant, because the coding agent comes with it.

Which should you choose?

• Choose GitHub Copilot if you want the lowest entry price, unlimited completions, a wide choice of models, and tight GitHub integration.
• Choose Cursor if you want an AI-native editor with cloud agents and are happy to lean on Grok and Composer models for most of your usage.
• Choose Claude Code if you prefer working in the terminal, want Anthropic’s models, and already pay for, or want, a Claude subscription.
• Choose Codex if you already pay for ChatGPT; it costs nothing extra and runs GPT-6 Sol from the Plus plan upward.

Many developers combine them. Copilot handles completions inside the IDE, while a subscription agent such as Claude Code or Codex takes the longer, multi-file jobs. With free tiers available for Copilot, Cursor and Codex, the practical advice is to try the same real task in two of them before committing to an annual plan.

Written byTheo Park

Theo Park runs the AI desk at Pandromeda. He follows model launches from the frontier labs and the open-weight community, tracks the assistants and developer tools built on them, and explains what each release changes on pricing, capability and safety. His reporting leans on primary sources: model cards, technical reports, API documentation and the companies' own announcements.

More from Theo ParkAll AI coverage