YearlyRank AI Tool2026AI tools listRanking
⌘K

How Claude Code, GitHub Copilot, and OpenAI Codex differ

· Comparison

Claude Code, GitHub Copilot, and OpenAI Codex are official coding agents that can work in an editor, a terminal, and related surfaces. This article compares what their vendors published, as of 16 November 2026, about capabilities, prices, and documented use cases, so you can match a tool to the way you already work.

The three products are already listed on YearlyRank AI Tool: Anthropic’s Claude Code, GitHub’s GitHub Copilot, and OpenAI’s Codex. Plan names and prices change; confirm current terms on the official sites before you buy or standardise on a tool. All three catalogue entries are still under methodology review, so this guide does not assign a score. See the methodology page.

Capabilities

Claude Code is Anthropic’s coding agent and assistant. The official product page calls it an agentic coding tool that understands a codebase, edits files, and runs commands. It invites users to hand Claude “a bug fix, test, or multi-day migration,” with permissions set by the user, and to steer and review work from the desktop, the terminal, an IDE, Slack, or the web. Official documentation (“Overview - Claude Code Docs”) also lists IDE extensions, a desktop app, and the browser. Anthropic names macOS, Linux, and Windows, and further entry points for Terminal, Web, iOS, Android, GitHub, VS Code, JetBrains, and Slack. A product FAQ says Pro, Max, Ultra, Team, and Enterprise users can use Claude Code in the Claude desktop app. The docs say most surfaces require a Claude subscription or an Anthropic Console account, and that the Terminal CLI, VS Code, and JetBrains integrations also support third-party providers.

GitHub Copilot is GitHub’s AI coding assistant and coding agent. The official product page is titled “GitHub Copilot · Your AI coding agent” and says Copilot works “from an open issue to reviewing and merging a pull request,” in the editor, the terminal, on the desktop, and on GitHub. GitHub Docs (“About GitHub Copilot”) add that it suggests code as you type, answers codebase questions, reviews changes, and works on assigned tasks, using the repository, issues, and pull requests already in that workflow. The docs group this into assistive use (suggestions you review and apply), agentic use (a multi-step goal, still subject to your review), customizations, and work with external agents, models, and tools, including MCP servers. The product FAQ lists Visual Studio Code, Visual Studio, JetBrains IDEs, XCode, Vim, Neovim, Zed, Eclipse, terminal/CLI, a desktop app, the web, mobile, Slack, and Teams.

Codex is OpenAI’s coding agent. The official page at chatgpt.com/codex is titled “Codex in ChatGPT | AI Coding Agents for Software Engineering.” OpenAI tells readers to use Codex “across ChatGPT, your editor, and the terminal, all connected by your ChatGPT account,” and lists Codex in ChatGPT, Codex cloud, a Codex IDE extension, and the Codex CLI. The same page says Codex is included in a ChatGPT plan, that it is available in the ChatGPT app, that work can continue in reusable cloud environments after you close a laptop, and that Codex can automatically review pull requests. An OpenAI.com announcement URL returned HTTP 403 on 16 November 2026, so this article uses chatgpt.com/codex rather than that announcement.

Claude CodeGitHub CopilotOpenAI Codex
VendorAnthropicGitHubOpenAI
What the vendor calls itAI coding agent and assistant; agentic coding tool for the terminal and IDEAI coding assistant and AI coding agentCoding agent for software engineering
Official surfaces namedDesktop, terminal, IDE, Slack, web; macOS, Linux, Windows; also Terminal, Web, iOS, Android, GitHub, VS Code, and JetBrains. Docs also list IDE extensions, a desktop app, and the browser.Editor, terminal, desktop, and GitHub, plus the IDEs and chat apps listed aboveChatGPT, the editor, and the terminal, tied to a ChatGPT account; Codex in ChatGPT, Codex cloud, a Codex IDE extension, the Codex CLI, and the ChatGPT app
What official pages say it doesUnderstands a codebase, edits files, runs commands; users can assign a bug fix, test, or multi-day migration and set permissionsSuggests code, answers codebase questions, reviews changes, and works on assigned tasks; can move from an issue to a pull request inside GitHubHandles coding tasks in ChatGPT, the editor, and the terminal; official page also describes cloud tasks that continue after you close a laptop, and automatic pull-request review
Official entry pointsanthropic.com/claude-code (also loaded as claude.com/product/claude-code); docs overviewgithub.com/features/copilot; About GitHub Copilotchatgpt.com/codex. openai.com/index/codex returned HTTP 403.
YearlyRank pageClaude CodeGitHub CopilotOpenAI Codex

Pricing

The figures below are what official pages stated on 16 November 2026 (Asia/Shanghai). They are not a quote or an invoice.

Claude Code. Official docs say most surfaces require a Claude subscription or an Anthropic Console account. On the product page, the Team plan was $20 per seat / month if billed annually, for teams of 2 to 150, with self-serve seat management and up to $500 per month in pooled API credits. Enterprise included everything in Team, plus Claude Security, usage and ROI analytics, role-based access, spend limits, and custom data controls; no Enterprise list price appeared in the retrieved HTML. An Individual list price was also not in the retrieved product-page HTML; see the vendor’s current pricing page. A FAQ also described “fast mode” as a research-preview configuration for Opus 5.5, priced at $8 / $40 per million tokens, on consumption-based plans and via usage credits on subscription plans.

GitHub Copilot. The product page listed Free at $0 USD (2,000 completions per month; access to Haiku 4.5, GPT-5 mini, and more; Copilot Fusion Standard; no credit card required); Pro at $10 USD per user / month (Copilot in the cloud and code review, unlimited completions and next-edit suggestions, access to third-party agents including Claude Code and Codex, model selection, and $15 monthly total credits); Pro+ at $39 USD per user / month (premium models including Opus, audit logs, a higher included-usage allowance, $70 monthly total credits, and Copilot Fusion Standard, High, and Max); and Max at $100 USD per user / month (priority access to new models and features, a higher included-usage allowance, and $200 monthly total credits). GitHub stated that 1 AI credit equals $0.01 USD, that chat and agent work use credits, and that completions and next-edit suggestions do not. Student, Business, and Enterprise list prices were not in the retrieved product-page HTML; see the vendor’s current pricing page.

Codex. OpenAI listed Plus, Pro, and Business and said Codex is included in the ChatGPT plan. Business was described as a shared workspace with admin controls and flexible pricing for teams using Codex across repositories, for 2+ seats, billed annually, “unlimited subject to abuse guardrails.” Dollar amounts for those plans were not in the retrieved HTML; see OpenAI’s current Codex page for current plan terms.

Use cases

Inline help while you type. GitHub’s documentation puts assistive use first: Copilot suggests code as you type, and you review and apply each suggestion. Completions and next-edit suggestions, on the plans above, do not consume GitHub AI Credits. Claude Code and Codex are presented first as agents you assign work to, although both also integrate with editors.

Assigning a task and reviewing the result. All three vendors describe this pattern. Anthropic asks you to hand Claude a bug fix, test, or migration and to set permissions. GitHub Docs describe agentic use in which Copilot may research a repository, propose a plan, edit files, and prepare work for your review — “you remain responsible for reviewing and approving.” OpenAI describes giving Codex engineering problems such as refactors and new features, then following progress and reviewing results, including automatic pull-request review.

GitHub-centred work, the terminal, and the cloud. Copilot is built into GitHub issues, pull requests, and repository automations, according to GitHub Docs, which describe a change that can begin as an issue and return as a pull request in the same place. Claude Code officially lists GitHub as a surface. Codex’s official page says it can review pull requests and keep cloud environments per repo. All three pages also name a CLI or terminal surface and an editor or IDE extension. OpenAI describes work that continues when you close a laptop; GitHub’s Pro plan text includes “Copilot in the cloud.” Claude Code includes a web surface. For chat and mobile, Claude Code lists iOS, Android, and Slack; GitHub lists mobile, Slack, and Teams; OpenAI says Codex is available in the ChatGPT app.

Buying for a team. Use the official plan text in the Pricing section, then request a current quote. Anthropic’s Team price is sourced; its Enterprise list price is not in the retrieved HTML. GitHub publishes four consumer prices and refers separately to Business and Enterprise, where admins set usage limits. OpenAI describes ChatGPT Business as a shared workspace with admin controls, billed annually for two or more seats, and says Codex is included in the ChatGPT plan.

How to choose

This closing advice follows the comparison above; it does not rank the three tools.

If your daily loop is GitHub issues and pull requests, start with GitHub’s “About GitHub Copilot” page and the plan table on github.com/features/copilot. As of 16 November 2026, the Pro plan text also includes access to Claude Code and Codex as third-party agents inside Copilot.

If you want a vendor-described agent you can steer from a terminal, an IDE, the web, or a phone, with a documented Claude or Console account requirement, read Anthropic’s product page and docs overview. The Team price above is sourced. For Individual and other plan prices, see the vendor’s current pricing page.

If you already have a ChatGPT plan and want the same agent in ChatGPT, an editor, and the terminal — including the cloud workflow OpenAI describes — start at chatgpt.com/codex and confirm which plan you have. List prices were not in the retrieved HTML; see the current Codex page.

Re-check the official pages before you decide, because surfaces and allowances are not fixed. YearlyRank lists each product separately and explains the evaluation method on the methodology page. This comparison is editorial background, not an official score.

Sources

Retrieved on 16 November 2026 (Asia/Shanghai):

  1. Anthropic, “Claude Code by Anthropic | AI coding agent and assistant,” https://www.anthropic.com/claude-code (also loaded as https://claude.com/product/claude-code). Product description, surfaces, Team and Enterprise plan text, desktop-app FAQ, and fast-mode FAQ.
  2. Anthropic, “Overview - Claude Code Docs,” https://docs.anthropic.com/en/docs/claude-code/overview (also loaded as https://code.claude.com/docs/en/overview). What Claude Code is, surfaces, account requirements, and the third-party provider note for Terminal, VS Code, and JetBrains.
  3. GitHub, “GitHub Copilot · Your AI coding agent,” https://github.com/features/copilot. Description, supported IDEs and surfaces, Free / Pro / Pro+ / Max prices and inclusions, GitHub AI Credits, and the Student-plan mention.
  4. GitHub Docs, “About GitHub Copilot,” https://docs.github.com/en/copilot/about-github-copilot/what-is-github-copilot. Assistive and agentic use, GitHub workflow integration, customizations, and external agents.
  5. OpenAI, “Codex in ChatGPT | AI Coding Agents for Software Engineering,” https://chatgpt.com/codex. Surfaces, inclusion in ChatGPT plans, Plus / Pro / Business plan names, cloud workflow, and pull-request review. Dollar amounts were not in the retrieved HTML.

Not used as a readable official source: https://openai.com/index/codex/ (HTTP 403 on 16 November 2026).

Related pages: Claude Code, GitHub Copilot, OpenAI Codex, and methodology.