Claude vs Codex limits, side by side
ClaudeCodex
Independent comparison. Not affiliated with Anthropic or OpenAI.
Official Claude Pro/Max and Codex (ChatGPT) plan limits, side by side. Quotes are from vendor plan, pricing, and help pages (plus named staff posts on our related guides). The dated limit reset log is LimitWatch's own list. Where a figure is unpublished, we write not stated. No third-party reporting or ranking.
Short answer
- Price band (individual paid): Claude Pro $20/mo; Max $100 or $200/mo. ChatGPT Plus (Codex) $20/mo; Pro $100, $200 or $500/mo; Go $8/mo. Numbers below are as published.
- Five-hour window: Claude Pro and Max have a session limit that resets every five hours. Codex Plus and Standard Business have a five-hour limit with estimated local messages; Pro plans currently have no five-hour limit.
- Weekly limits: Claude Pro and Max have weekly limits; size not stated. Codex: weekly limits "may also apply"; size not stated.
- Absolute size in messages or tokens: not stated for Claude. Codex publishes only estimates for Plus/Business local messages per five hours.
- After the cap: Claude offers usage credits (API rates) when enabled, and occasional limit resets. Codex Plus/Pro can buy credits (price per credit not stated); OpenAI also offers occasional/banked resets. OpenAI Support does not reset limits on request.
- Resets logged here: see the AI usage limit resets log (16 OpenAI rows, 1 Anthropic row as of that page).
Side-by-side table
| Topic | Claude (Anthropic) | Codex (OpenAI / ChatGPT) |
|---|---|---|
| Individual plan prices (monthly billing) | Pro $20; Max $100; Ultra $200 (claude.com/pricing; Max help) | Go $8; Plus $20; Pro $100 / $200 / $500; Business $25/user (Codex pricing) |
| Five-hour / session window | Yes — session limit resets every five hours (Pro and Max; Ultra: no 5-hour limit for verified accounts). Absolute size: not stated. | Plus and Standard Business: five-hour window with estimated local messages (not fixed limits). Pro: no five-hour limit. Absolute fixed size: not stated. |
| Weekly limits | Yes (Pro, Max and Ultra). Size: not stated. | "May also apply." Size: not stated. |
| Model-specific / extra weekly (Claude) | Max and Ultra: separate weekly limit for Fable (size not stated). Pro: no separate Fable weekly called out on the Claude Code help page. | No Fable-equivalent model-specific weekly called out on Codex pricing. |
| Included monthly API credits | Max $100/month; Ultra $200/month on Claude pricing. Cannot be used in Claude apps or Claude Code (including usage credits). Free/Pro: none. | Not an included monthly API credit allotment on Codex pricing; Plus/Pro can purchase additional credits (price per credit not stated). |
| Published multiples / estimates | Max = 5× Pro usage every five hours and every week; Ultra = 10× Pro weekly usage. Message/token counts: not stated. | Plus/Business local msgs per 5h (estimates): GPT-6 Astra 5–45; GPT-6.1 Sol 15–160; GPT-6 Sol 15–150; GPT-6.1 Terra 35–285; GPT-6 Luna 350–3,000. Staff: Pro 20X ≈ 20× Plus weekly; Pro 500 announced as 25× Plus. Absolute Pro sizes: not stated. |
| Shared across products? | Yes — Claude and Claude Code share the same plan limits. | Codex shares an allowance/credit pool with ChatGPT Work, ChatGPT for Excel and Workspace Agents when those features are on the plan. ChatGPT pricing (2026-10-27): Free "Limited Codex access"; Plus "Expanded Codex usage"; Pro "Longer Codex and ChatGPT Work sessions"; compare grid Free/Go Limited then Expanded (Pro cell blank in the capture). |
| What uses the allowance faster (official) | Research sessions "may consume tokens more quickly" (help center). Relative cost of models/agents in messages: not stated. | Fast mode: 2.5× included / 2× credits; Ultrafast: 8× included / 6× credits; images ~3–5× faster; cloud tasks may use more than local. Auto-review checks do not count. |
| After you hit the cap | Wait for reset; upgrade; or enable usage credits (standard API rates). Minimum top-up: not stated. | Plus/Pro can buy credits (price per credit not stated). OpenAI Support will not reset limits on request. |
| Limit-reset policy (official help) | Occasional resets for eligible plans; you choose when to use one. Who/how often: not stated. | Occasional one-time or banked resets; eligibility varies; future resets not guaranteed. A full banked reset refreshes 5-hour and weekly windows and moves the weekly reset date. |
| Recent announced limit change (from official pages) | 2026-09-22: higher five-hour limits on Pro/Max/Team/seat Enterprise; size of increase not stated. Same day: a savable rate-limit reset for subscription users. Entry: Anthropic (Claude): Five-hour usage limits increased on Pro, Max, Team and seat-based Enterprise plans at the Claude Opus 5.5 launch; subscription users also got a saveable rate limit reset. Weekly limits are not mentioned in the announcement. (2026-09-22). | No cut announced on the Codex pricing page we cite. Pro 500 (25× Plus) announced 2026-09-29 on @OpenAI. Staff have announced multiple paid-user resets; see the reset log. |
| Staff replies to "limits draining faster" | Anthropic Claude Code staff (Oct 2026) replied publicly that limits have not changed; asked for /usage or /feedback. Not a formal product announcement. | OpenAI Codex staff (Aug 2026) said they do not change limits without saying so; also described cache-hit and other inefficiencies for some users. Details of any fix: not stated in the captured posts. |
Claude: how limits are described
Paid plans have two types of usage limits: a session limit that resets every five hours, and weekly limits that reset once every week.
These limits are shared across Claude and Claude Code, meaning all activity in both tools counts against the same limits.
Additional usage limits may apply to manage capacity and ensure fair access to all users. For example, we may limit your usage in other ways, such as weekly and monthly caps or model and feature usage, at our discretion.
Max gives you 5x the usage of Pro every 5 hours and every week.
Size of any Claude weekly or five-hour limit in messages or tokens: not stated. More detail: How Claude's weekly usage limits work · Claude limits draining faster? What's confirmed.
Codex: how limits are described
The estimates below show local messages per five-hour period for Plus and Standard Business.
These estimates are not fixed message limits
Pro plans currently have no five-hour limit.
Weekly limits may also apply.
Weekly limit size for any Codex plan: not stated. Fixed (non-estimate) five-hour limits: not stated. More detail: Codex limits: what's confirmed.
Resets compared
LimitWatch keeps a dated log of usage-limit resets announced on official pages or by named staff. As built into /guides/ai-limit-resets/: 16 OpenAI (Codex / ChatGPT Work) rows and 1 Anthropic row. Each row quotes the source verbatim. Resets known only from news or user posts are left out.
Limit resets are given occasionally to eligible plans, and set your usage limits back to full when you choose to use one.
OpenAI may occasionally provide eligible users with a one-time reset of Codex or ChatGPT Work usage limits. Eligibility, the limits affected, and any redemption or expiration requirements may vary by offer. Future resets are not guaranteed.
Eligibility, frequency and whether a given outage triggers a reset: not stated beyond what each help article and the dated log already quote.
User reports (not confirmation)
People on X compare Claude Max and Codex meters. Those are user reports, not vendor statements. LimitWatch only lists user reports on /reports/ after human review, with neutral wording, until an official source confirms them. This comparison page does not treat report posts as sources for any number in the table.
What the sources don't say
- Which product gives more coding usage for the same dollars: not stated in comparable units.
- Claude five-hour or weekly size in messages/tokens: not stated.
- Codex weekly limit size for any plan; fixed five-hour caps; Free/Go Codex numbers beyond "Limited": not stated.
- Price per Claude usage-credit top-up minimum, or per Codex credit: not stated.
- Size of the 2026-09-22 Claude five-hour increase: not stated.
- Whether either vendor reduced subscription limits in October 2026: no announcement found on the official pages cited here; Anthropic staff replies say no changes; OpenAI staff said they do not change Codex limits silently.
Sources
- https://chatgpt.com/pricing (checked 2026‑10‑30)
- https://claude.com/pricing (checked 2026‑11‑06)
- https://developers.openai.com/codex/pricing (checked 2026‑10‑30)
- https://help.openai.com/en/articles/11369540-using-codex-with-your-chatgpt-plan (checked 2026‑10‑24)
- https://support.claude.com/en/articles/11049741-what-is-the-max-plan (checked 2026‑10‑30)
- https://support.claude.com/en/articles/11145838-use-claude-code-with-your-pro-max-or-ultra-plan (checked 2026‑11‑02)
- https://support.claude.com/en/articles/11647753-how-do-usage-and-length-limits-work (checked 2026‑11‑04)
- https://support.claude.com/en/articles/12429409-manage-usage-credits-for-paid-claude-plans (checked 2026‑11‑06)
- https://support.claude.com/en/articles/17007452-what-is-a-limit-reset (checked 2026‑10‑30)
- https://support.claude.com/en/articles/8325606-what-is-the-pro-plan (checked 2026‑10‑30)
- https://www.anthropic.com/claude-opus-5-5 (checked 2026‑10‑22)
- https://x.com/OpenAI/status/2104993967985381673 (checked 2026‑10‑26)
- https://x.com/bcherny/status/2113141937645769087 (checked 2026‑10‑25)
- https://x.com/lydiahallie/status/2113366380842254746 (checked 2026‑10‑25)
- https://x.com/thsottiaux/status/2090675027670978569 (checked 2026‑10‑26)
- https://x.com/thsottiaux/status/2094254532020818191 (checked 2026‑10‑26)
Also used: LimitWatch's own AI usage limit resets log (each row there cites a primary source).
New: ChatGPT Plus vs Pro after the Oct 29 doubling · Pro 200 limit cut reversed.
See also: Claude limits draining faster? What's confirmed · Codex limits: what's confirmed · AI usage limit resets log · What each AI coding plan gets you · How Claude's weekly usage limits work · AI status tracker. Something wrong? Dispute it; corrections are dated and appended.