Claude Code vs Codex: Which to Use for Client Builds

Claude Code (from Anthropic) and Codex (from OpenAI) are AI coding agents that read your project, edit files and run commands. Both run in a terminal, IDEs, a desktop app and the cloud, and support MCP, skills and hooks. The main differences are the models, the plans and workflow details.

This comparison is for people building tools and apps for clients. It's based on each vendor's own docs and pricing pages, checked October 15, 2026. Both products change often, so check the linked pages before you decide. Stride AI Academy teaches both and isn't affiliated with Anthropic or OpenAI.

Quick comparison

	Claude Code	Codex
Made by	Anthropic	OpenAI
Terminal CLI	Yes	Yes
IDE	VS Code (and Cursor) extension, JetBrains plugin	IDE extension
Desktop app	Yes (Code tab in the Claude app)	Yes (ChatGPT desktop app)
Web and cloud	claude.ai/code, cloud sessions	ChatGPT on the web, Codex Cloud
GitHub	GitHub Actions, code review	GitHub integration and a GitHub Action
Instructions file	CLAUDE.md (can also read AGENTS.md)	AGENTS.md
Config file	settings.json files	~/.codex/config.toml
MCP	Yes	Yes
Skills	Yes	Yes
Hooks	Yes	Yes
Plugins	Yes	Yes
Subagents	Yes	Yes

Where each one runs

Claude Code runs in the terminal, in VS Code and JetBrains IDEs, in the Claude desktop app, on the web at claude.ai/code and in the Claude mobile app. You can also mention @Claude in Slack to route a task. On macOS, Linux and WSL, the docs give this install command (Claude Code overview):

curl -fsSL https://claude.ai/install.sh | bash

Codex runs in the ChatGPT desktop app, ChatGPT on the web and mobile, the Codex CLI, the Codex IDE extension, Codex Cloud and GitHub (Codex docs).

Both let you start a long task in the cloud and check on it later, which is handy when a build takes a while.

How you pay

Claude Code is included with Claude's paid plans. Anthropic's pricing page lists Pro at $20 a month billed monthly ($17 a month billed annually), Max from $100 a month, and Team seats from $25 a month billed monthly. The Free plan doesn't include Claude Code (claude.com/pricing). You can also use it with an Anthropic Console account billed by API usage.

Codex is included with ChatGPT plans. OpenAI's Codex pricing page lists Free and Go ($8 a month) with lighter usage, Plus at $20 a month, Pro from $100 a month and Business at $25 per user a month billed monthly. You can also use Codex with an API key in the CLI, SDK or IDE extension, billed at API rates, but without the cloud features (Codex pricing).

Checked against Anthropic and OpenAI pricing pages, October 15, 2026. Usage limits differ by plan and change, so check current limits for the plan you're considering.

Instructions and config files

Both agents read a markdown file of project instructions at the start of a session. This is where you put the project's conventions: stack, commands, folder layout and rules like "never edit the migrations folder."

• Claude Code reads CLAUDE.md. If your repo already has an AGENTS.md, Claude Code can read that instead (Claude Code memory docs).
• Codex reads AGENTS.md. It checks a global file in ~/.codex, then walks from the project root down to your current directory, using at most one file per directory. An AGENTS.override.md takes priority over AGENTS.md at the same level. The combined content is capped at 32 KiB by default (Codex AGENTS.md docs).

If you use both tools on the same repo, keep one AGENTS.md as the source of truth so both agents get the same rules.

Settings live in different places. Claude Code uses settings.json files (~/.claude/settings.json for you, .claude/settings.json for a project). Codex uses ~/.codex/config.toml, with project-level .codex/config.toml for trusted projects.

Extensibility: MCP, skills, hooks and plugins

MCP servers connect the agent to outside tools such as GitHub, a database or Stripe. Adding one looks like this:

# Claude Code
claude mcp add --transport http stripe https://mcp.stripe.com

# Codex
codex mcp add vercel --url https://mcp.vercel.com

Claude Code also supports project-scoped servers in a .mcp.json file you can commit, so everyone on the project gets the same tools (Claude Code MCP docs). Codex configures servers in config.toml under [mcp_servers.<name>] (Codex MCP docs). The Codex example above comes from Vercel's MCP setup page.

Skills package a repeatable workflow (instructions plus optional scripts) the agent can load when relevant. Both tools support them. In Claude Code, a skill is a folder with a SKILL.md file in .claude/skills/ or ~/.claude/skills/. See how to create Claude Code skills.

Hooks run your own commands at fixed points, such as formatting a file after every edit or blocking edits to .env. Both tools document hooks. See the Claude Code hooks tutorial.

Plugins bundle skills, hooks and MCP servers so you can share a setup across projects. Both tools support them.

Which to use for client work

Both can build the same kinds of client projects: dashboards, internal tools, web apps, API integrations and automations. Practical ways to choose:

• You already pay for one. If you have ChatGPT Plus or a Claude Pro plan, start with the agent it includes and see whether it handles your projects.
• Try both on the same task. Give each the same spec on a copy of the same repo and compare the result, the number of corrections you made and the time it took. Your own projects are a better test than published benchmarks.
• Team setup. If clients or teammates already use one of the two, matching it makes handover easier.
• Use both. Plenty of builders do. A shared AGENTS.md keeps both agents on the same rules, and you can use one to review the other's work.

Whichever you pick, the skills that matter are the same: writing a clear spec, breaking work into small steps, reviewing diffs, testing and deploying safely.

Learn both

The Stride AI Coding course teaches Claude Code and Codex for building custom tools and apps for clients, written for non-programmers, with weekly live calls to review your builds. Many of the free guides cover specific Claude Code and Codex setups, and the Claude Code community page explains what members work on.

FAQ

Is Codex better than Claude Code?

Neither is better for every task. Both are capable coding agents with similar features, and results depend on the task, the model and how clearly you describe the work. Test both on your own project with the same spec and compare.

Does Codex use Claude Code?

No. Codex is OpenAI's coding agent and uses OpenAI models. Claude Code is Anthropic's coding agent and uses Claude models. They're separate products, though both support open standards like MCP and AGENTS.md.

Can I use Claude Code and Codex on the same project?

Yes. Keep shared project instructions in AGENTS.md, which Codex reads and Claude Code can read. Commit your work often with Git so you can compare or undo changes from either agent.

Do I need to know how to code to use them?

You don't need to write the code yourself, but you do need to describe what you want clearly, review what the agent changed and test that it works. Knowing basic Git and how your project runs helps a lot.