Claude CodevsCodex CLI
Generated from the two spec rows. Green marks the better value where a spec has a clear direction. Everything else is just different.
Claude Code
Anthropic · Terminal agent
#3MCP
Panel
7.03 spec wins
- Reliability
- 7.2
- Usefulness
- 8.0
- Cost
- 5.5
- Longevity
- 7.2
“Included with Pro at $20, which is how you learn what the Max plan is for.”
Codex CLI
OpenAI · Terminal agent
#1OSSMCP
Panel
7.14 spec wins
- Reliability
- 7.3
- Usefulness
- 7.0
- Cost
- 6.8
- Longevity
- 7.2
“Apache-2.0, so you can read every line except the one that lets you pick a model that is not OpenAI's.”
Spec by spec
| Spec | Claude Code | Codex CLI |
|---|---|---|
| Architecture | ||
| Category | Terminal agent | Terminal agent |
| Runs | local, cloud | local |
| Platforms | macos, linux, windows | macos, linux, windows |
| Context window | 200k tokens, 1M on select models | not documented |
| Protocols | ||
| MCP client | Yes | Yes |
| MCP server | Yes | No |
| Capabilities | ||
| Runs terminal commands | Yes | Yes |
| Multi-file edits | Yes | Yes |
| Git operations | Yes | Yes |
| Browser control | Yes | NoThe shared Codex codebase carries a browser-use configuration with Chrome DevTools Protocol access, but OpenAI's documentation states browser is not available in the Codex CLI or the IDE extension; it runs in ChatGPT on the web and in the desktop app. The CLI's --search flag adds live web search only. |
| Sandboxed execution | Yes | Yes |
| Multi-agent orchestration | Yes | Yes |
| Headless / CI mode | Yes | Yes |
| Models | ||
| Backbone | Claude | OpenAI GPT-5.x, GPT-5.x-Codex |
| Bring your own model | YesThrough Amazon Bedrock, Google Cloud's Agent Platform, Microsoft Foundry, or any LLM gateway your organisation runs, with model IDs pinned per deployment. | Yes |
| Local models | NoClaude Code routes only to Claude models. Anthropic states it does not support routing Claude Code to non-Claude models through any gateway, and gateway model discovery drops every model ID that does not contain "claude" or "anthropic", so a local Ollama or LM Studio endpoint is not a supported setup. | Yes |
| Cost | ||
| Pricing model | mixed | mixed |
| Starts at | $20/mo | $8/mo |
| Free tier | No | Yes |
| Bring your own key | YesYour own Anthropic API key, or your own AWS, Google Cloud or Azure account. | Yes |
| Openness | ||
| Open source | No | Yes |
| License | proprietary | Apache-2.0 |
| GitHub stars | 148,889 | 127,563 |
Which one would each critic pick
| Critic | Claude Code | Codex CLI | Pick |
|---|---|---|---|
| El Juez | — | — | not enough reviews |
| El Amigo | 7.8 | 8.0 | Codex CLI — If you already pay for ChatGPT, this is the terminal agent you have and it is a good one; open source, MCP, headless, and pointed at OpenAI until you configure it otherwise. |
| El Crítico | 6.8 | 7.0 | Codex CLI — A sandbox mode you choose once is the only barrier between the agent and your shell, the free tier is a taste, and the defaults point at one vendor. |
| El Profesor | 6.8 | 6.3 | Claude Code — A capable architecture whose claims are supported by no benchmark on record; the design is principled and unmeasured. |
| La Inversora | 8.5 | 8.3 | Claude Code — Anthropic owns the model, the distribution and the invoice, which is the rarest kind of moat in this category. |
| La Jefa | 7.0 | 7.3 | Codex CLI — Business at $20 a seat with SAML SSO and no default training, Enterprise for SCIM and audit logs; if the org is on ChatGPT already, this is a line item, not a project. |
| El Hacker | 5.0 | 5.8 | Codex CLI — Apache-2.0 source I can actually read and fork, MCP in a TOML file, AGENTS.md as the contract, and an --oss flag that finally lets my own box onto the model list. |
Picks are derived from each critic's own scores. Humans vote on matchups on the duels page.