Codex vs Claude Code in 2026: pricing, limits and features

Quick answer: both are agentic coding tools that come with a consumer subscription. Codex is part of ChatGPT plans: Plus at $20/month, Pro from $100 ($100, $200 or $500), and Business at $20 per user per month billed annually. Claude Code is part of paid Claude plans: Pro at $20/month, Max at $100 or $200, and Team from $20 per seat. Both use rolling five-hour windows plus weekly limits, with some exceptions, and both let you continue on paid credits. Which is better depends on your work. Neither company's docs name a winner, and we won't either.

Key facts

• Entry price for full coding use: $20/month on both (ChatGPT Plus, Claude Pro).
• Top individual plan: ChatGPT Pro $500/month, or Claude Max 20x $200/month.
• Pro (ChatGPT) currently has no five-hour limit for Codex. Claude Max still has five-hour sessions.
• Codex's recommended model: GPT-6.1 Sol. Anthropic's default suggestion for most workloads: Claude Opus 5.5.
• Check usage: /status in the Codex CLI, /usage in Claude Code.

Plans and prices

Codex (from OpenAI's pricing page):

• Free ($0) and Go ($8/month): GPT-6 Luna at Standard speed in the desktop app, subject to rollout.
• Plus ($20/month): Codex on the web, in the CLI, in the IDE extension and on iOS, plus cloud features like automatic code review and Slack. Includes GPT-6.1 Sol and GPT-6 Luna.
• Pro (from $100/month: $100, $200 or $500): everything in Plus. Ultrafast mode is on Pro $500.
• Business ($20/user/month billed annually, $25 monthly, 2+ users) and Enterprise/Edu (contact sales).
• API key: CLI, SDK or IDE extension billed at API rates, with no cloud features.

Claude Code (from claude.com/pricing and Anthropic's help center):

• Free: Claude Code isn't included.
• Pro ($20/month, or $17/month billed annually): includes Claude Code.
• Max 5x ($100/month) and Max 20x ($200/month): 5x or 20x Pro's usage per session, plus $100 or $200 a month in API credits for the Claude Platform.
• Team: Standard seats $20/month billed annually ($25 monthly), Premium seats $100 ($125). Team covers 2 to 150 seats.
• API: Claude Code can also run on an Anthropic Console account at API rates.

How usage limits work

Codex: OpenAI publishes estimated local messages per five-hour window for Plus and Standard Business. On Plus, for example, it estimates 15 to 160 GPT-6.1 Sol messages. Weekly limits may also apply, and the estimates aren't fixed limits. Pro plans currently have no five-hour limit. Codex and ChatGPT Work share usage. Once you hit your limits, Plus and Pro users can buy credits.

Claude Code: Pro and Max have a five-hour session limit and a weekly limit, and Max also has a separate weekly limit for Fable, per Anthropic's help center. Usage is shared with Claude on the web, desktop and mobile. Anthropic doesn't publish message estimates for Claude Code. Paid plans can turn on usage credits to keep working.

Where they run and how to install

• Codex: the ChatGPT desktop app, web, CLI, IDE extension and iOS. Install the CLI with curl -fsSL https://chatgpt.com/codex/install.sh | sh on macOS/Linux, or with npm install -g @openai/codex, or brew install --cask codex. Sign in with ChatGPT on first run.
• Claude Code: terminal, VS Code and Cursor, JetBrains, a desktop app, and the web at claude.ai/code, including the Claude mobile apps. Install with curl -fsSL https://claude.ai/install.sh | bash on macOS/Linux/WSL, or with brew install --cask claude-code or winget install Anthropic.ClaudeCode.
• Both read project instruction files. Codex uses AGENTS.md. Claude Code uses CLAUDE.md, and per its docs it can read an existing AGENTS.md instead.

Models

OpenAI's models page recommends GPT-6.1 Sol for complex coding and agentic work and Luna for focused, repeatable tasks. It describes GPT-6 Astra as its most capable model. GPT-5.5 retires from ChatGPT, Work and Codex on 14 October 2026. Anthropic's models overview suggests starting with Claude Opus 5.5 for most workloads and using Claude Fable 5.1 for demanding reasoning and long-horizon agentic work. Sonnet 5.5 and Haiku 5.5 cover faster and cheaper work.

What outside benchmarks say

We stick to independent numbers and attribute them. On 12 October, Artificial Analysis reported that Claude Code with Sonnet 5.5 scored 68 on its Coding Agents Index, with Opus 5.5 at 66. It reported cost per task on its Intelligence Index of $3.26 for GPT-6 Astra and $0.72 for GPT-6.1 Sol. These are one lab's figures at one point in time.