Comparison · under IB-CODE-2026.2
Claude Code vs OpenAI Codex CLI
| Claude Code | OpenAI Codex CLI | |
|---|---|---|
| Vendor | Anthropic | OpenAI |
| Pricing | Claude Pro $20/mo; Max $100/mo (5x) or $200/mo (20x); API at $3 / $15 per Mtok (Sonnet 4.6 in/out) | Included in ChatGPT Pro $200/mo; or raw API at GPT-5.5 rates |
| Default model | claude-sonnet-4-6 (Opus 4.6 on flagship tasks) | GPT-5.5 (GPT-5.5-mini available for cost-sensitive runs) |
| CLI surface | ✓ | ✓ |
| Eval status | published · read | queued |
One side evaluated, one side pending
Claude Code has been evaluated under IB-CODE-2026.2: read the Claude Code evaluation →
The other side (OpenAI Codex CLI) is currently queued . When that eval lands, this page will update to a full head-to-head comparison.
About each tool
Claude Code
Claude Code is Anthropic's agentic coding CLI, GA as of early 2026. Ranked #1 most-loved in the Pragmatic Engineer dev survey (Feb 2026, 15K respondents). Default model is Sonnet 4.6 with Opus 4.6 routing on flagship tasks. The eval anchor for Indie Bench because it's the tool being operated to run …
OpenAI Codex CLI
The Codex name has a long history — the original 2021 Codex API (code-davinci-002) was deprecated in March 2023. The current Codex CLI is a complete re-architecture: GPT-5.5 backed, with Goal Mode for multi-hour autonomous runs, native subagents, MCP server support, plugin marketplace, browser verif…