2026-11-05Compare
ChatGPT + Codex Share One Allowance in December: Stay on Plus, Upgrade Pro, Buy Credits, or Move to API?
Open Settings → Usage on your ChatGPT account tonight. You already pay for Plus, Pro, or Business, and you just lived through the October 29 Work and Codex double. OpenAI Help now says that starting in December, ChatGPT and Codex get a shared usage allowance on those paid plans. Tonight’s bill question is simple: stay on Plus, upgrade Pro, buy credits for spikes, or move heavy agentic work to API.
[Image: Diagram of ChatGPT and Codex merging into one December shared usage pool | alt="December shared usage pool for ChatGPT and Codex on Plus Pro Business"]
Key takeaway: Instant chat will not draw from your usage. Advanced chat, Codex, and ChatGPT Work–style agentic tasks will. OpenAI has not published new per-plan token budgets or an exact December calendar day. Treat this as a billing redesign, not a free expansion.
This guide covers consumer and team ChatGPT subscriptions—not OpenAI Platform API rate cards. Live list prices sit on Deepapi’s Subscriptions hub; verify against chatgpt.com/pricing if anything drifts.
| Plan | Monthly price | In December shared pool? | Instant chat | Best if … | Reconsider when … |
|---|---|---|---|---|---|
| Free | $0 | Not named in December Help change | Within Free limits | Occasional experiments | You need Codex Cloud or paid headroom |
| Go | $8/mo | Not named in December Help change | Within Go limits | Light daily chat | You need advanced Codex Cloud work |
| Plus | $20/mo | Yes — ChatGPT + Codex shared | Exempt | Instant + light advanced chat; occasional Codex | Settings → Usage walls mid-week on agentic tasks |
| Pro 100 | $100/mo (monthly only) | Yes | Exempt | Solo builders who burn Plus; Ultrafast not included | Work/Codex pool drains before month-end |
| Pro 200 | $200/mo (monthly only) | Yes | Exempt | Full-time multi-file agent loops; Ultrafast not included | Pro 100 bars stay low after a full cycle |
| Pro 500 | $500/mo (monthly only) | Yes | Exempt | Ultrafast needed (this tier only) + max included usage | You rarely use Ultrafast |
| Business | $25/user/mo (annual often $20/user/mo; min 2) | Yes | Exempt | Small teams needing workspace controls | Most work is personal Plus/Pro |
| API + router | Pay-per-token | Outside ChatGPT subscription pool | N/A | Batch, multi-model, schedulable agents | You still need native canvas/voice daily |
What OpenAI Said About December Shared Usage
OpenAI’s Help article Understanding ChatGPT and Codex usage limits (updated days before this November 5, 2026 research pass) states:
| Policy point | What buyers should hear |
|---|---|
| Timing | Starting in December (no exact day on the Help page) |
| Plans named | Plus, Pro, and Business |
| What merges | Usage allowance shared across ChatGPT and Codex |
| Why | Bringing agentic capabilities of Work into main ChatGPT |
| Meter | Tokens; advanced models and complex tasks generally use more |
| Exempt | Instant chat will not draw from your usage |
| Dashboard | Combined remaining usage and reset times in Settings → Usage |
| At the limit | Wait for reset, upgrade your plan, or buy credits for eligible features |
| Soft reassurance | Expects most users won’t wait for resets more often than now—without publishing new budgets |
A companion Help page on using Codex with your ChatGPT plan already describes shared allowances among Codex, ChatGPT Work, ChatGPT for Excel, and Workspace Agents. The December change folds more of main ChatGPT (outside Instant chat) into that logic—no merge multiplier is published. Early-November TechRepublic, The New Stack, and TechRadar coverage matches: named plans, Instant exempt, absolute budgets unpublished.
This Is Not the July Codex Merger
Do not confuse this page with Deepapi’s GPT-5.6 Codex merger guide. That July 2026 story answered where Codex lives in the ChatGPT desktop/Work product. This December story answers which tokens come from which pool. UI consolidation is not a shared usage allowance.
How the October 29 Double Changes the Math
On October 29, OpenAI doubled Work and Codex–related allowances on Pro tiers and canceled a planned Pro 200 cut. See Deepapi’s post-double audit: ChatGPT Pro after the double—$100, $200, or $500?. This page does not rewrite that guide.
December math is interaction, not rehash:
- The October double made the Pro pool larger for Work and Codex workloads.
- The December shared allowance puts more ChatGPT surfaces onto the same token budget Codex already draws from.
- A larger pool serving more surfaces is not automatically more headroom than last month.
Keep the double guide for Pro-tier selection. Use this guide for Plus, Pro, Business, credits, and API choices under one shared pool.
December Bill Ladder at a Glance
List prices reflect Deepapi’s live Subscriptions snapshot for November 5, 2026. If official pricing changes, trust chatgpt.com/pricing and the live hub.
[Image: Plus vs Pro vs Business vs API ladder for December shared ChatGPT Codex pool | alt="December ChatGPT Codex bill ladder Plus Pro Business API"]
Notes: Codex appears across plans including Free and Go, but Codex Cloud is for eligible Plus, all Pro tiers, Business, Enterprise, Healthcare, and Education—not Free or Go—per OpenAI’s Codex plan Help. Free/Go were not listed as December shared-pool targets in the primary Help article.
Stay on Plus ($20)
Keep Plus when:
- Most of your day is Instant chat, which will not draw from the shared allowance.
- Advanced reasoning and Codex are occasional spikes, not all-day loops.
- You will not pay five to twenty-five times Plus pricing for unpublished December budgets.
- Settings → Usage after a normal week still shows comfortable headroom.
Plus remains the default for generalists who need conversational help, light coding assist, and rare agentic runs. “December sounds scary” is not an upgrade trigger.
Cross-check a single $20 seat against our fall 2026 subscription bill ladder and best AI subscription under $20 guides.
When to Upgrade Pro (100 / 200 / 500)
Upgrade when the meter—not the headline—forces the decision.
Pro 100 ($100/mo): Plus repeatedly throttles agentic Codex or advanced ChatGPT Work–style tasks, and Instant chat is not the bottleneck. Monthly only. Ultrafast not included.
Pro 200 ($200/mo): Continuous terminal or multi-file agent loops would still exhaust Pro 100 after the October double. Ultrafast not included.
Pro 500 ($500/mo): You need Ultrafast, included on this tier only among the consumer Pro ladders above. Do not buy Pro 500 solely because December is shared.
Because absolute December token budgets are unpublished, capture a baseline week in Settings → Usage before leaping tiers. Count reset waits. Note whether Instant-heavy days leave the pool untouched. Then price the gap. For Pro-only post-double nuance, use Pro 100 vs 200 vs 500 after the double.
Buy Credits vs Upgrade vs Wait
Help lists three actions at the limit: wait for reset, upgrade your plan, or buy credits for eligible features.
- Wait when collisions are rare.
- Buy credits for sharp spikes when most months still fit the included pool.
- Upgrade when you hit the wall weekly and credits would become a second recurring bill.
Credit unit prices are outside this article’s verified facts—check in-product flows and official Help. Credits are a spike valve, not a fix for the wrong base plan.
Move Heavy Work Off the Shared Pool
If expensive work is batchable, multi-model, or schedulable, stop forcing it through the ChatGPT–Codex shared allowance.
Route background extraction, CI summaries, overnight refactors, and eval harnesses through developer endpoints. Compare live token cards on Model API Price. For fall 2026 routing math, see the fall 2026 API bill guide.
Claude Code and OpenRouter-style gateways can absorb repetitive agent loops so your ChatGPT seat stays focused on Instant chat plus selective advanced sessions. Inside Codex, shrink context before buying a higher seat—see reducing Codex token usage with context compaction.
Moving spend to API does not cancel a ChatGPT seat if you still need canvas, voice, or native Work UI daily—it only clears workloads that should not compete for one shared consumer pool.
Tonight’s ChatGPT–Codex Bill Checklist
[Image: Tonight’s ChatGPT Codex shared pool bill checklist | alt="Tonight’s ChatGPT–Codex bill checklist before December shared allowance"]
- Open Settings → Usage and screenshot remaining usage plus reset times as your pre-December baseline.
- Split last week into Instant chat (exempt) versus advanced chat / Codex / ChatGPT Work (pooled).
- Count how often you already wait on resets after the October 29 double.
- Challenge any assumption that “doubled = December will feel looser.”
- Decide if pain is a rare spike (credits) or a weekly wall (upgrade).
- If eyeing Pro 500, confirm you need Ultrafast—not just a bigger shared pool.
- Move one batch workload to API this week; price it on Model API Price.
- Apply context compaction so Codex burns fewer tokens per task.
- Recheck live Plus / Pro / Business prices on Subscriptions.
- Re-read Help on ChatGPT and Codex usage limits.
- Skim Today for clarifications after this guide’s research date (November 5, 2026, Asia/Singapore).
FAQ
What exactly changes in December for ChatGPT and Codex usage?
Starting in December, Plus, Pro, and Business plans’ usage allowance will be shared across ChatGPT and Codex as OpenAI brings agentic Work capabilities into the main ChatGPT experience. Usage is measured in tokens. Instant chat will not draw from your usage. Combined remaining usage and reset times appear in Settings → Usage.
Which plans are affected—and which are not clearly included?
OpenAI’s Help article names Plus, Pro, and Business. Free and Go are not framed as the same December shared-allowance change in that article. Enterprise and education packaging follow their own contracts.
Does Instant chat count against the shared allowance?
No. Instant chat will not draw from your usage. Advanced models and complex agentic tasks generally consume more tokens from the shared pool.
Is this the same as the July 2026 Codex–ChatGPT product merger?
No. The July merger is about product placement and Work/Codex UX. This December change is metering: one shared usage allowance across ChatGPT and Codex on named paid plans. See the Codex merger guide.
How does the October 29 Pro Work/Codex double interact with a December shared pool?
The double increased Work/Codex headroom on Pro tiers. The December shared pool expands which ChatGPT surfaces draw from related metering. Larger allowance plus broader consumption can cancel out—audit Settings → Usage. Details: Pro usage doubled guide.
Has OpenAI published the new token budgets?
No—as of November 5, 2026, Help does not publish new per-plan absolute token budgets or an exact December day. Soft reassurance about reset frequency is not a numeric guarantee.
Should I stay on Plus, jump to Pro, buy credits, or leave for API / Claude Code?
Stay on Plus if Instant chat dominates and agentic spikes are rare. Upgrade Pro when Settings → Usage shows steady walls after the double. Buy credits for occasional peaks. Move batch and multi-model work to API, Claude Code, or a router when those jobs should not compete inside the consumer shared pool.
Where do I check live remaining usage and live plan prices?
Check Settings → Usage in ChatGPT after the change lands. Check consumer list prices on Subscriptions and chatgpt.com/pricing. For developer token cards, use Model API Price.
The Bottom Line on Tonight’s Bill
The December shared pool does not automatically cancel your ChatGPT subscription, and it does not automatically justify Pro 500. It reallocates how Plus, Pro, and Business buyers spend one allowance across ChatGPT and Codex—while Instant chat stays outside that draw.
Measure first. Keep Plus when the meter is calm. Step into Pro 100 or Pro 200 when agentic walls are real. Reserve Pro 500 for Ultrafast. Use credits for spikes. Push heavy schedulable work onto API routes on Model API Price. Continue with the Pro double sibling, the Codex merger explainer, the subscription bill ladder, live Subscriptions, and Today.
Deepapi: independent AI consumer intelligence—impartial, sources linked, not affiliated with any AI company.
Current prices are on the price table.