Claude Code (Anthropic) vs OpenAI Codex CLI
Claude Code (Anthropic) vs OpenAI Codex CLI: side-by-side comparison of two agent ides & coding harnesses systems — architecture, taxonomy, license, pricing, MCP/A2A support, and direct edges.
Cost & capability
| Claude Code (Anthropic) | OpenAI Codex CLI | |
|---|---|---|
| Capability band | frontier | frontier |
| Capability composite | 88 | 80 |
| Cost tier | mid | free |
| $/Mtok input | — | 0 |
| $/Mtok output | — | 0 |
| Use cases | Code Generation Focused, Long Running Session, Scoped Agentic | Code Generation Focused, Scoped Agentic, Long Running Session |
Where they differ (16)
Rows where both sides have data and the values disagree — the shortlist of dimensions that actually distinguish these two systems.
| Claude Code (Anthropic) | OpenAI Codex CLI | |
|---|---|---|
| Capability composite | 88 | 80 |
| Cost tier | mid | free |
| Use cases | Code Generation Focused, Long Running Session, Scoped Agentic | Code Generation Focused, Scoped Agentic, Long Running Session |
| Type | Terminal-native agentic coding harness | Terminal-native agentic coding harness (OpenAI) |
| Created | 2024-10 | 2025-04 |
| Latest release | v2.1.310 (2026-10-23) | v0.165.0 (2026-10-27) |
| License | proprietary — © Anthropic PBC, use subject to Anthropic Commercial Terms (repo public, not OSS) | Apache-2.0 |
| GitHub | anthropics/claude-code — ~153k stars, last commit 2026-10 | openai/codex — ~128k stars, last commit 2026-10 |
| Pricing | No free tier; included in Pro ($20/mo, $17 annual), Max (from $100/mo), Team ($25/seat/mo, $20 annual), Enterprise; or API pay-per-token via Console | CLI free (Apache-2.0); usage via ChatGPT plans (Free/Go limited, Plus $20/mo, Pro $100–200/mo, Business $25/user/mo) or API-key per-token billing |
| Funding | Anthropic (parent) — $65B Series H at $965B post-money (2026-05-28); S-1 filed for a proposed IPO (2026-10-26) | OpenAI (parent) — $40B Series I 2025-03 $300B val |
| Backend storage | local CLAUDE.md + ~/.claude state + git | local ~/.codex/ + git |
| Deployment | Both (self-hosted + managed cloud) — local CLI/IDE/desktop client plus cloud sessions on web; models via Anthropic API (or Bedrock/Vertex) | Self-hosted only — local CLI (npm/Homebrew); model calls go to OpenAI via ChatGPT sign-in or API key |
| API surface | CLI (Node.js); subagents; hooks; MCP | CLI (Rust); MCP support |
| MCP | yes — MCP client (claude mcp add; stdio and HTTP servers) | yes — MCP client (mcp_servers in ~/.codex/config.toml); documented in Codex MCP docs |
| Optimised for | Anthropic Claude as a long-running coding agent; MCP + skills + hooks composable surface | OpenAI o-series + Codex model as long-running coding agent |
| Anti-fit | Anthropic-only (no GPT, Gemini); terminal-only — no GUI | OpenAI-model-only; CLI-only |
At a glance
| Claude Code (Anthropic) | OpenAI Codex CLI | |
|---|---|---|
| Section | Agent IDEs & coding harnesses | Agent IDEs & coding harnesses |
| Tier | T1 | T1 |
| Type | Terminal-native agentic coding harness | Terminal-native agentic coding harness (OpenAI) |
| Created | 2024-10 | 2025-04 |
| Latest release | v2.1.310 (2026-10-23) | v0.165.0 (2026-10-27) |
| License | proprietary — © Anthropic PBC, use subject to Anthropic Commercial Terms (repo public, not OSS) | Apache-2.0 |
| GitHub | anthropics/claude-code — ~153k stars, last commit 2026-10 | openai/codex — ~128k stars, last commit 2026-10 |
| Pricing | No free tier; included in Pro ($20/mo, $17 annual), Max (from $100/mo), Team ($25/seat/mo, $20 annual), Enterprise; or API pay-per-token via Console | CLI free (Apache-2.0); usage via ChatGPT plans (Free/Go limited, Plus $20/mo, Pro $100–200/mo, Business $25/user/mo) or API-key per-token billing |
| Funding | Anthropic (parent) — $65B Series H at $965B post-money (2026-05-28); S-1 filed for a proposed IPO (2026-10-26) | OpenAI (parent) — $40B Series I 2025-03 $300B val |
| Backend storage | local CLAUDE.md + ~/.claude state + git | local ~/.codex/ + git |
| Deployment | Both (self-hosted + managed cloud) — local CLI/IDE/desktop client plus cloud sessions on web; models via Anthropic API (or Bedrock/Vertex) | Self-hosted only — local CLI (npm/Homebrew); model calls go to OpenAI via ChatGPT sign-in or API key |
| API surface | CLI (Node.js); subagents; hooks; MCP | CLI (Rust); MCP support |
| Multi-tenancy | per-user (terminal-local) | per-user (terminal-local) |
| MCP | yes — MCP client (claude mcp add; stdio and HTTP servers) | yes — MCP client (mcp_servers in ~/.codex/config.toml); documented in Codex MCP docs |
| A2A | searched not found | searched not found |
| OpenTelemetry | searched not found | searched not found |
| Optimised for | Anthropic Claude as a long-running coding agent; MCP + skills + hooks composable surface | OpenAI o-series + Codex model as long-running coding agent |
| Anti-fit | Anthropic-only (no GPT, Gemini); terminal-only — no GUI | OpenAI-model-only; CLI-only |
Pros & cons
Claude Code (Anthropic)
Pros: Reference harness for Claude; richest extensibility surface in the field (subagents, skills, hooks, MCP) — the others copy it.
Cons: Anthropic-only model lock; CLI-first UX is a barrier vs IDE chat panes; pricing rides Anthropic API usage rather than seat tiers.
OpenAI Codex CLI
Pros: OSS Apache-2.0 means hackable by community; first-party OpenAI distribution + auto-routing to Codex / o-series.
Cons: OpenAI-only model lock; younger than Claude Code with less ecosystem (skills/hooks); naming collision with deprecated old Codex API and ChatGPT "Codex" cloud agent.