Skip to content
AI Productivity Insights
Tutorial · November 2026

ChatGPT and Codex Will Share One Usage Pool — How to Make It Last

A shared agentic pool is already live for Codex, ChatGPT Work and related features, while regular Chat is still separate. Starting in December, Plus, Pro and Business plans will share one token allowance across ChatGPT and Codex. Instant chat will not draw from that December pool. Here's where to check remaining usage, what burns it, and habits that keep coding time available.

Tutorial, ChatGPT + Codex. Bright blue illustration with the title 'ChatGPT joins Codex in one pool in December' and the line 'Plus, Pro, Business: one ChatGPT + Codex token pool — Instant doesn’t count'. Three tiles read 'When: Starting in December', 'Unit: Tokens' and 'Check: Settings → Usage'. On the right, a chat icon and a terminal icon feed arrows into a glowing pool with a usage meter.

Introduction

OpenAI Help (re-checked 15 Nov 2026; the article still showed an update from about 13 days earlier) says that starting in December, Plus, Pro and Business plans will share one token-based usage allowance across ChatGPT and Codex as Work’s agentic capabilities move into main ChatGPT. Instant chat will not draw from that usage. That December change is not the first shared meter. Official Help, checked 15 Nov 2026, already says that on plans with shared allowances or credits, Codex, ChatGPT Work, ChatGPT for Excel, and Workspace Agents use a shared allowance and credit pool when those features are available. Official personal-plan credits Help, also checked 15 Nov 2026, adds that for Plus and Pro, Codex, ChatGPT Work, Excel, and PowerPoint can share the same agentic usage allowance when those features are available on your plan, and that Word uses that allowance too. Regular Chat is still metered separately today. Enterprise, Edu, Free, and Go are outside the December Plus, Pro and Business sentence.

The short version

  • Today: on plans with shared allowances or credits, Codex, ChatGPT Work, ChatGPT for Excel, and Workspace Agents already use a shared pool when those features are available. Official credits Help adds that, for Plus and Pro, Excel and PowerPoint can share the same agentic usage allowance, and that Word uses that allowance too. Regular Chat is still separate.
  • Starting in December: Plus, Pro and Business only — no exact calendar day. ChatGPT and Codex will share one token allowance. Not Enterprise, Edu, Free, or Go.
  • Instant: will not draw from that December usage; heavier models and complex tasks generally use more tokens.
  • Check: Settings → Usage for remaining allowance and reset times.

1. Two Changes — Don’t Mix Them Up

A vertical blue-spine timeline, re-checked 15 Nov 2026. 29 Oct is already live: official release notes increased Work and Codex usage for Pro and Business Premium; Pro 100 and Pro 500 doubled; the Pro 200 reduction was canceled and lower-allowance seats were doubled to the same limit. Starting in December, marked as not yet in effect: Plus, Pro and Business will share one ChatGPT + Codex token allowance. Instant chat will not draw from your usage. Codex remains a separate desktop surface, and official Help has not described a Chat + Work merge as a general rollout. No published per-plan token caps; use Settings → Usage.
Figure 1: The 29 Oct Pro and Business Premium increase is already live; the December ChatGPT + Codex token pool is a further, separate change.

Separately — and already live as of 29 Oct 2026 — official ChatGPT release notes say OpenAI increased the Work and Codex usage included in Pro and Business Premium plans. Pro 100 and Pro 500 now include twice their previous Work and Codex allowance. OpenAI also canceled the Pro 200 reduction announced on 29 Sep: subscribers who already had the higher allowance keep it, and Pro 200 subscribers who were on the lower allowance get twice that previous allowance, bringing them to the same limit. That change is not the December shared-pool design. (Our 30 Oct walkthrough covers the earlier Pro-limits story. Official release notes for 28 Oct 2026 also describe promotional API-organization credits for eligible Pro and Business plans; those are separate from ChatGPT usage credits and are not the December pool.)

OpenAI says it is preparing carefully and expects most users won’t need to wait for resets more often than they do now. That is OpenAI’s expectation, not a measured guarantee.

2. What Draws From the Shared Pool

Two cards and a bottom note, re-checked 15 Nov 2026. Today, already live: on plans with shared allowances or credits, Codex, ChatGPT Work, ChatGPT for Excel, and Workspace Agents use a shared allowance and credit pool when those features are available; Word uses that allowance too; Excel and PowerPoint can share it on Plus and Pro; regular Chat is still metered separately. 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 that usage; not Enterprise, Edu, Free, or Go; not yet in effect as of 15 Nov 2026. A bottom note: official Help still treats Chat and Work as separate, with Codex a separate desktop view; a 1 Nov subscriber email, as reported by BleepingComputer, described a small-group test, not a Help announcement of a general merge.
Figure 2: Today’s Work/Codex agentic pool is already live; the December ChatGPT + Codex token pool is a further change. Instant stays off that December meter.
SurfaceTodayStarting in December (Plus, Pro, Business)
Instant chat—No — will not draw from that usage
Regular ChatStill metered separately from the current Work/Codex poolYes — part of the shared ChatGPT + Codex token allowance (Instant excluded)
Codex, ChatGPT Work, ChatGPT for Excel, and Workspace AgentsAlready share an allowance and credit pool when those features are available on the plan. Word uses that allowance too; Excel and PowerPoint can share it on Plus and Pro.Codex is in the December ChatGPT + Codex token allowance. Official Help describes the December change as sharing one token allowance across ChatGPT and Codex as Work’s agentic capabilities move into main ChatGPT.

Usage for the December pool “will be measured in tokens.” How much you use depends on the model and the task: more advanced models and complex tasks generally use more. Official Help still does not publish per-plan weekly or monthly token caps for the December pool, so don’t invent budgets — read Settings → Usage on your own account. When you are signed in with a ChatGPT account, official Codex-with-plan Help says auto-review safety checks are free and do not count toward plan usage limits; other Codex activity still does. Regular Chat Voice has separate limits and does not consume Codex usage; task work started through Voice in Work or Codex on desktop uses the same shared agentic allowance, and connected Voice time can be metered separately.

3. Walkthrough: Keep the Pool for Coding

Four editorial step cards, re-checked 15 Nov 2026: baseline Settings → Usage, noting remaining usage and the reset time; prefer Instant so brainstorming and short Q&A stay off the December token pool; shorter Codex threads and the smallest model that finishes the job; at the wall, wait, upgrade, or buy credits from Settings → Usage. A caption points to official credits Help for expiry rules and some rates.
Figure 3: Baseline first, then Instant for light work, then watch model choice on Codex.

Step 1. Baseline Settings → Usage

Open Settings → Usage and note remaining usage plus the reset time before December arrives. That snapshot makes it easier to spot whether the shared pool feels tighter after the change — without guessing from memory.

Step 2. Prefer Instant for light work

Use Instant for brainstorming and short Q&A so the shared pool stays available for Codex and heavier agentic work. OpenAI’s wording is clear: Instant chat will not draw from your usage.

Step 3. Shorter threads, clearer tasks

For coding, keep threads short and tasks specific. Avoid re-pasting huge logs when a link or file attach works. Pick the smallest model that finishes the job — advanced models generally burn more tokens.

Step 4. Keep Codex as its own desktop surface

Official Help, re-checked 15 Nov 2026, still says ChatGPT includes Chat and Work as separate experiences, and that Codex remains a separate view in the ChatGPT desktop app. A 1 Nov subscriber email, as reported by BleepingComputer, said OpenAI would test a combined Chat + Work experience with a small group; that test has not been described on official Help as a general rollout. Official Help has not said that everyone gets one merged surface in December.

Step 5. If you hit the wall

Wait for the reset, upgrade the plan, or buy credits to continue using eligible features. Official Help still does not publish per-plan weekly or monthly token caps for the December pool. Credit rates and expiry rules are on OpenAI’s credits and rate-card pages — including Settings > Usage in ChatGPT, or Usage & Billing in the Codex app — not in a blog post that invents USD pack prices. Official personal-plan credits Help, updated 15 Nov 2026, says credits are a pay-as-you-go add-on after included limits, for eligible features including Codex, ChatGPT Work, Word, Excel, and PowerPoint; included usage is used first; credits can be used automatically if you keep using Work or Codex after the included allowance runs out. Eligible Plus and Pro users can buy credits without changing plans. Credits expire 12 months from purchase; credits purchased in Japan after 13 Nov 2026 expire in 180 days; unused credits do not roll over. That page publishes credits-per-1-million-tokens rates for ChatGPT for Word, Excel/Sheets, and PowerPoint, and points Work and Codex rates to the individual pricing guide. Official rates for Word, Excel/Sheets, and PowerPoint, in credits per 1 million tokens, are: GPT-5.6 Luna 5 input / 0.5 cached input / 30 output; GPT-5.6 Terra 50 / 5 / 300; GPT-5.6 Sol 100 / 10 / 500; GPT-6 Astra 250 / 25 / 1,250. Model availability can vary by feature and region. This post does not add USD pack prices or weekly token caps.

↑ All steps

4. Common Misconceptions

The December ChatGPT + Codex token pool is not the same as today’s already-live Work/Codex agentic pool, and it is not the same as the 29 Oct Pro and Business Premium Work and Codex increase. Instant will not draw from the December usage under OpenAI’s current wording — that is not a license to invent “unlimited forever” claims beyond what Help says. Official Help still does not publish per-plan weekly or monthly token caps for the December pool; credit rates and expiry rules are on OpenAI’s credits pages, not as USD pack prices invented here. Official Help still treats Chat and Work as separate, and Codex stays a separate desktop view; a small-group Chat + Work test was described in a 1 Nov subscriber email as reported by BleepingComputer, not as a general Help rollout.

5. What Teams Should Do Before December

If your team already leans on Codex for daily coding and ChatGPT for planning, treat December as a further metering change, not a feature drop, and not the first time Codex has shared a pool with Work. Ask everyone to open Settings → Usage once, note the reset window, and decide which work stays on Instant. For Pro and Business Premium seats, remember the 29 Oct Work and Codex increase already raised those included ceilings — including the official Pro 200 nuance that the 29 Sep reduction was canceled and lower-allowance seats were doubled to the same limit. The December change is separate: Plus, Pro and Business plans will then share one token allowance across ChatGPT and Codex.

Do not invent weekly token budgets for Plus, Pro or Business in planning docs. Point teammates at Help and at their own Usage panel. If someone routinely hits the wall on eligible features, the documented options are wait for the reset, upgrade the plan, or buy credits — using official credit rates and expiry rules, not USD pack prices invented in a blog post. Enterprise and Edu billing is a separate credit or USD token arrangement and is outside this Plus/Pro/Business tutorial.

Conclusion

A shared agentic pool is already live for Codex, ChatGPT Work and related features, while regular Chat is still separate. Starting in December, Plus, Pro and Business will share one token allowance across ChatGPT and Codex. Instant stays off that December meter. Check Settings → Usage, keep Instant for light work, watch model choice on Codex, and treat the 29 Oct Pro and Business Premium increase as a separate, already-shipped change. Official Help still does not publish per-plan token caps for the December pool; use official credits Help for rates and expiry, and re-read those pages before you set team budgets.

Resources

Know a teammate who burns Codex all day and Instant all night? Share this with them before December.

Newsletter

Get AI productivity insights delivered

Practical AI guides, tutorials, and fresh insights in your inbox.