Codex vs Claude Code: What Each $20 Plan Gets You Before the Usage Limit
Prices, five-hour and weekly windows, and every way past a limit, from OpenAI’s and Anthropic’s own pages.The standard plan for each costs $20 a month. ChatGPT Plus includes OpenAI Codex, and Claude Pro includes Claude Code. Both meter you in a five-hour window with a weekly cap on top, and both…
Annons
Annons
Prices, five-hour and weekly windows, and every way past a limit, from OpenAI’s and Anthropic’s own pages.The standard plan for each costs $20 a month. ChatGPT Plus includes OpenAI Codex, and Claude Pro includes Claude Code. Both meter you in a five-hour window with a weekly cap on top, and both let you pay to keep going.They differ on the free tier, on the next plan up, and on what happens to the task you’re in the middle of when the meter runs out.Google’s autocomplete for “codex vs claude code” now suggests “usage limits” and “cost,” so that’s what we cover here. We checked every number on the vendors’ own pages on October 20, 2026. We haven’t run the two agents side by side on one project, so this post sticks to plans and limits and makes no claims about code quality.What each plan costs, and where the agent is includedChatGPT FreePrice (Oct 20, 2026): $0Coding: Codex, limited, in the desktop app (subject to rollout)ChatGPT GoPrice: $8/monthCoding: Codex for lightweight tasks, desktop app (subject to rollout)ChatGPT PlusPrice: $20/monthCoding: Codex on the web, CLI, IDE extension and iOS, plus automatic code review on GitHubChatGPT ProPrice: $100, $200 or $500/monthCoding: Everything in Plus, and no five-hour limit for nowClaude FreePrice: $0Coding: No Claude CodeClaude ProPrice: $20/month, or $17/month billed annually ($200 up front)Coding: Claude Code in the terminal and in VS Code, Cursor and JetBrains IDEsClaude Max 5xPrice: $100/monthCoding: 5x Pro’s usage per five-hour sessionClaude Max 20xPrice: $200/monthCoding: 20x Pro’s usage per five-hour sessionSources: OpenAI’s Codex pricing page, Anthropic’s pricing page and its Max plan article.If you want to try before paying, only Codex has a $0 option. Claude’s pricing table lists Claude Code on Pro, Max 5x and Max 20x, with no entry for Free.How the five-hour window worksOpenAI publishes rough numbers. For Plus, its pricing page estimates 15 to 160 local messages per five-hour period on GPT-6.1 Sol, and 350 to 3,000 on the lighter GPT-6 Luna. The page calls these estimates and says cloud tasks may use more of your allowance than local messages.OpenAI’s help center adds when the clock starts. A new five-hour window begins when you send your first message after the previous window ends.Anthropic publishes no message counts. Its pricing FAQ says every plan’s limits “reset on a rolling five-hour session window,” that Pro gives at least 5x Free’s usage per session, and that Max gives 5x or 20x Pro’s.Both pools are shared. On Claude, chats on the web, desktop and mobile apps draw from the same pool as Claude Code. On ChatGPT, Codex shares its allowance with ChatGPT Work, and local messages share it with cloud chats.The weekly cap on topAnthropic’s support article on Claude Code with Pro or Max says both plans have a five-hour session limit and a weekly limit. OpenAI’s pricing page says weekly limits “may also apply,” and its help center says that where both apply, you need allowance left in both to continue.So a long week can stop you with hours left in a session window. ChatGPT Pro currently drops the five-hour limit, but OpenAI says the plan still has an included allowance, so the weekly meter keeps counting. Claude Max keeps both windows.When the limit lands mid-taskBoth vendors let a running step finish. OpenAI’s pricing page says that if you hit your limit “during an active turn, the agent will be able to continue working on that turn, subject to fair use limits.”Anthropic calls its version the wrap-up allowance. Claude Code shows “Usage limit reached · wrapping up” and finishes the response in progress, and the extra usage counts toward your weekly limit. On Pro it applies up to once per weekly period, on Max each time you hit a five-hour limit, and Anthropic says it’s still rolling out.Neither lets you send a new message after the cap. Commit your work before you start a long refactor, and check the meter first. In the Codex CLI that’s /status; in Claude Code, /usage shows plan usage bars for subscribers.Every way past a limitWaitCodex: Your usage dashboard shows the reset timesClaude Code: Settings > Usage shows where you standPay per use on your planCodex: Plus and Pro can buy ChatGPT creditsClaude Code: Turn on usage credits, billed at standard API ratesMove up a planCodex: Pro from $100/monthClaude Code: Max 5x at $100, Max 20x at $200Use an API keyCodex: Local chats at API rates, without cloud features such as GitHub code reviewClaude Code: A Claude Console account with API creditsBuy a resetCodex: Eligible personal Plus and Pro accounts, depending on billing countryClaude Code: Not offered on the pages we readOpenAI has also handed out free resets since September, both automatic ones and “banked” ones you apply yourself. Its help center treats them as promotions: a reset restores your allowance but doesn’t raise your plan’s limits. Don’t plan a deadline around the next one.What drains the allowance fasterFast mode on Codex. Included usage burns at 2.5x the Standard rate.Switching models mid-pool. OpenAI’s help center says it doesn’t restore anything you’ve used.Extra context on every message. OpenAI’s tips name long AGENTS.md files and MCP servers you aren’t using.A stray API key in Claude Code. If ANTHROPIC_API_KEY is set in your environment, Claude Code bills the API instead of your Pro or Max plan.Max’s monthly API credits. Anthropic says they can’t be used in Claude Code.How we’d pickIf you have $0, start with Codex on ChatGPT Free and see whether the desktop app’s limited access is enough. If you already pay for ChatGPT Plus or Claude Pro, use the agent you already have before buying the other.If you keep hitting the five-hour wall, the two $100 plans solve it differently. ChatGPT Pro removes the five-hour limit for now and leaves the weekly one. Claude Max 5x keeps both windows and gives you five times Pro’s usage in each five-hour session.Two $20 plans cost $40 and give you two separate pools, each with its own five-hour window. That’s an option if you hit limits a few days a month rather than every day, and it’s less than either $100 tier.If you’re also choosing an editor to run these agents in, our Claude Code vs Cursor comparison covers where each one fits.FAQIs Codex free? ChatGPT Free and Go ($8/month) include limited Codex in the desktop app, subject to rollout. Codex on the web, CLI and IDE extension starts with Plus at $20 a month.Is Claude Code free? No. Claude’s pricing page lists Claude Code on Pro ($20/month, or $17/month billed annually) and on Max ($100 or $200/month).How do Codex usage limits reset? Plus has a five-hour window that starts with your first message after the last one ended, and weekly limits may also apply. ChatGPT Pro currently has no five-hour limit. Your usage dashboard and /status in the CLI show the reset times.How do Claude Code usage limits reset? Claude plans use a rolling five-hour session window, and Pro and Max add a weekly limit. The pool is shared with your Claude chats.What happens when I hit the limit in the middle of a task? Codex finishes the active turn, subject to fair use. Claude Code may finish the response in progress under its wrap-up allowance, which is still rolling out. On both, a new message waits for the reset unless you pay for credits.This story is published under the Generative AI publication. Connect with us on LinkedIn and follow Zeniteq to stay in the loop with the latest AI stories. Let’s shape the future of AI together!Codex vs Claude Code: What Each $20 Plan Gets You Before the Usage Limit was originally published in Generative AI on Medium, where people are continuing the conversation by highlighting and responding to this story.