From December, ChatGPT and Codex share one usage allowance on Plus, Pro, and Business
source re-read 2026-11-02
Why this label: Primary claims are OpenAI’s own Help Center wording on the shared December allowance, Instant-chat exclusion, token metering, and Settings → Usage. We used the public Help Center page via search/snippet (Cloudflare challenge blocked a full HTML scrape from the box). Community X reactions are secondary color only. We reproduced nothing.
- Tried
- Keep separate mental budgets for ChatGPT chat work and Codex / agentic coding on a Plus, Pro, or Business plan.
- Broke
- OpenAI documents that from December those plans’ usage allowance is shared across ChatGPT and Codex as agentic Work capabilities move into the main ChatGPT experience. Usage is measured in tokens; more advanced models and complex tasks use more. Instant chat does not draw from the shared allowance. Hitting the limit means waiting for reset, upgrading, or buying credits where available — OpenAI says it expects most users will not need to wait more often than today, but heavy Codex users who also chat hard will feel one pool.
- Cause
- confirmed OpenAI Help Center “Understanding ChatGPT and Codex usage limits”: Starting in December, Plus, Pro, and Business plans’ usage allowance will be shared across ChatGPT and Codex; measured in tokens; Instant chat will not draw from usage; combined remaining usage and reset times appear in Settings → Usage.
- suspected Public X posts (late Oct–early Nov 2026) frame this as punishing power Codex users whose chat and coding will compete. That framing is community reaction, not a published multiplier or cut size from OpenAI on the page we cited.
- Fix
- Before December, check Settings → Usage and OpenAI’s usage-limits help page for how your plan will show the combined pool. Pace Codex runs knowing chat (except Instant) and coding share the same allowance; use credits or a plan change if you routinely exhaust both. Do not assume Support can restore a separate Codex bucket — the help page describes one shared allowance.
- Prevention
- Track Codex and non-Instant ChatGPT as one budget after December. Prefer Instant chat for light Q&A when you need to preserve coding headroom, and schedule heavy agent runs early in a reset window.