COMPARISON · Last verified: September 2026

Codex vs Claude Code

OpenAI's and Anthropic's coding agents, compared on verified pricing, surfaces, models and features.

Codex vs Claude Code: official demos

Watch each tool in action: official demos for Codex and Claude Code, side by side, before you decide.

Codex

►

Claude Code

►

Official videos via YouTube (OpenAI; Anthropic), embedded for reference. ToolChase does not host or claim these videos.

What is Codex?

OpenAI Codex is OpenAI's coding agent. It reads your code, makes changes, runs commands and tests, reviews diffs and can open pull requests. It runs as the open-source Codex CLI on macOS, Linux and Windows; as an IDE extension for VS Code and compatible editors such as Cursor and Windsurf (now Devin Desktop), with Xcode and JetBrains IDEs offering their own Codex integrations; as a Codex view in the ChatGPT desktop app, into which OpenAI merged the standalone Codex app on July 9, 2026; and as Codex cloud, which runs tasks in parallel in isolated environments and can start work from GitHub, GitLab (beta), Linear or Slack. Remote in the ChatGPT mobile app steers work on a connected Mac or Windows PC. Codex reads AGENTS.md project instructions and supports skills, plugins, MCP, hooks, subagents and scheduled tasks. It is included in every ChatGPT plan, from Free to Enterprise, or runs on an OpenAI API key at API rates.

What is Claude Code?

Claude Code is Anthropic's agentic coding tool. It reads your codebase, plans a change, edits files across the project, runs commands and tests, fixes what fails, and works with git: it stages changes, writes commit messages, creates branches and opens pull requests. It started in the terminal and now runs on surfaces that share one engine: the CLI on macOS, Windows and Linux, a VS Code extension that also installs in Cursor, a JetBrains plugin, the Code tab of the Claude desktop app, the web at claude.ai/code and the Claude app for iOS and Android. It reads CLAUDE.md, can also read AGENTS.md, and supports MCP, skills, plugins, hooks and subagents. It runs in GitHub Actions and GitLab CI/CD, takes tasks from Slack, and runs scheduled cloud routines, a research preview. It is included in every paid Claude plan or billed per token through the Anthropic API.

Key differences at a glance

The two agents overlap heavily. These are the differences that matter most:

• -- Price of entry: Codex is included in every ChatGPT plan, including Free and Go ($8/mo); Claude Code is not on Claude's Free plan and starts with Pro at $20/mo ($17/mo billed annually).
• -- Models: Codex's recommended models are GPT-6 Astra and GPT-5.6 Sol, Terra and Luna; Claude Code uses Claude models such as Fable 5.1, Opus 5 and Sonnet 5.
• -- Desktop app: Codex lives inside the ChatGPT desktop app, on every ChatGPT plan; Claude Code's desktop surface is the Code tab of the Claude app, on paid plans.
• -- Hosted cloud work: Codex cloud works with GitHub and GitLab (beta) repos; Claude Code's cloud sessions, a research preview, need GitHub to clone repos and open pull requests, so GitLab teams run Claude Code in GitLab CI/CD (beta).
• -- Pull request review: Codex reviews GitHub pull requests on Plus and higher plans; Anthropic's managed Code Review is a Team and Enterprise research preview billed on usage (about $15-25 per review on average), and any paid Claude plan can run reviews through GitHub Actions.
• -- Local safety: Codex sandboxes local commands by default; Claude Code starts in auto mode on Pro, Max and Team, where a classifier model reviews actions, and offers an optional OS-level Bash sandbox.
• -- Score: Claude Code rates 4.6/5 in our review; Codex rates 4.5/5.

Feature-by-feature comparison

How the two compare on daily-use capabilities, as each vendor documents them:

Pricing comparison

Codex comes with every ChatGPT plan; what changes is how much you can use. Free ($0) is pitched at "quick coding tasks" and Go ($8/month) at "lightweight coding tasks". Plus ($20/month) covers Codex on the web, in the CLI, in the IDE extension and on iOS, plus automatic code review and Slack, and OpenAI pitches it for "a few focused coding sessions each week". Pro starts at $100/month for 5x the Plus limits; OpenAI temporarily paused new sign-ups for the 20x tier on September 10, 2026. Business is $20 per user per month billed annually (2+ users) or $25 billed monthly, and Enterprise and Edu are quoted by sales. With an OpenAI API key, Codex bills at API rates, for example GPT-5.6 Sol at $4/$20 per million input/output tokens (promotional pricing OpenAI says runs at least through November 21, 2026), but without cloud features such as GitHub code review.

Claude Code is not on Claude's Free plan. It is included in Pro at $20/month billed monthly, or $17/month billed annually ($200 up front), and in Max from $100/month, which offers 5x or 20x Pro's usage and is billed monthly. Team (2 to 150 people) includes it on Standard seats at $20 per seat per month billed annually or $25 monthly, and on Premium seats, with 5x the usage, at $100 billed annually or $125 monthly. Enterprise is $20 per seat per month billed annually plus usage at API rates. Usage is shared with the rest of your Claude plan and resets on a rolling five-hour window, with weekly limits on paid plans. Through the Anthropic API, Sonnet 5 costs $2/$10 and Opus 5 $5/$25 per million input/output tokens.

Worked example. A 5-person team pays the same on either: ChatGPT Business or Claude Team Standard costs 5 × $20 = $100 a month billed annually ($1,200 a year), or 5 × $25 = $125 a month billed monthly. Both plans let you mix seat types, so giving one heavy user a Premium seat (5x the usage of a Standard seat) makes the bill 4 × $20 + $100 = $180 a month billed annually on either ChatGPT Business or Claude Team.

Bottom Line

Codex and Claude Code now overlap more than they differ: both run in a terminal, IDEs, a desktop app, the cloud and on your phone, with MCP, skills, hooks and subagents. Codex (4.5/5) is the cheaper way in, since every ChatGPT plan includes it, and it suits teams that want hosted cloud tasks on GitLab or managed GitHub pull request reviews in a $20 plan. Claude Code (4.6/5) is our pick for large autonomous changes such as refactors and migrations, and the natural choice if you already pay for Claude. From $20 a month up the main price tiers match, and Codex can import a Claude Code setup if you want to try both.

📚 Related