Starting in December 2026, OpenAI will merge the usage limits of ChatGPT and Codex on the Plus, Pro and Business plans into one shared allowance, measured in tokens, the units of text a model reads and writes. Ordinary Instant chat will not count against it. Advanced models, long tasks, Fast modes and Codex coding work will. The price does not change: Plus stays at $20 (about SAR 75) a month and Business at $20 per user billed annually. What changes is who runs out first. If the same employees use ChatGPT for daily work and Codex for code, decide before December which tasks deserve the advanced models.
What exactly changes in December
According to OpenAI's help page on ChatGPT and Codex usage limits, the change has four parts:
- One allowance: on Plus, Pro and Business, usage will be shared across ChatGPT and Codex, as OpenAI brings the agentic capabilities of ChatGPT Work into the main ChatGPT experience.
- Measured in tokens: how much you use depends on the model you choose and the task. More advanced models and complex tasks use more of the allowance.
- Instant chat is free of it: OpenAI states that Instant chat will not draw from your usage.
- One place to watch: the combined remaining usage and reset times will appear in Settings → Usage.
OpenAI has not published a start day within December, nor a token figure for each plan. It says it expects most users will not need to wait for their usage to reset more often than they do now. That is an expectation, not a guarantee.
Why this matters for a team, not just one user
Today, ChatGPT Work and Codex already share usage, according to OpenAI's official pricing page, while regular chat is counted separately. From December, the line between the two disappears for the paid plans. In practice, a developer who spends the morning running long Codex tasks may find less left in the afternoon for drafting a proposal in ChatGPT. An employee who uses an advanced model to summarize every email consumes allowance that Instant chat would have handled for free.
The model choice therefore becomes a cost decision. The pricing page notes that GPT-6.1 Sol is available in Work and Codex, not in Chat. For the differences between the models, see our article on GPT-6.1 Terra, Sol and Luna.
What drains the allowance faster
- Fast mode: uses included subscription usage at 2.5 times the normal rate.
- Ultrafast mode on GPT-6 Astra and GPT-6.1 Sol: 8 times the included subscription usage.
- Advanced models and long agentic tasks: more tokens per request than a quick question.
When the limit is reached, the help page lists three options: wait for the reset, upgrade the plan, or buy credits to keep using eligible features.
Current prices, for reference
- Plus: $20 a month (about SAR 75).
- Business: $20 per user per month billed annually for 2 or more users, or $25 (about SAR 94) billed monthly.
- Pro: $100, $200 or $500 a month.
If your company pays for Business or Pro, also check the free API credits that come with those plans, covered in our article on free ChatGPT Business API credits. API usage is billed separately from the subscription allowance.
Five steps before December
- List who on the team uses both ChatGPT and Codex on the same account. Those are the people most likely to hit the limit.
- Agree on a simple rule: Instant chat for quick questions, rewriting and short summaries; advanced models only for analysis, code and long documents.
- Turn off Fast and Ultrafast modes by default and keep them for urgent work.
- Open Settings → Usage in the first week of December and note how much each heavy user consumes in a normal day.
- Decide in advance who may buy credits or upgrade, and up to what amount, so the work does not stop while waiting for an approval.
The Origami view
A shared allowance is fair in principle: you pay for what you use. The risk for companies is invisible consumption, where one heavy task eats the budget of a whole working day. We see the same pattern when clients build AI into their own systems: the expensive part is rarely the subscription, it is using the strongest model for tasks that do not need it. When we build AI features for our clients, we route simple tasks to light models and keep the advanced ones for work that actually needs them, with a clear usage report. If you want to estimate your team's consumption or build an internal AI tool on a predictable budget, contact us or see our services.
Sources
- OpenAI Help Center — Understanding ChatGPT and Codex usage limits: the December change, Plus, Pro and Business plans, measurement in tokens, Instant chat excluded, Settings → Usage, and the options at the limit.
- OpenAI — ChatGPT pricing: Plus, Business and Pro prices, Work and Codex sharing usage, Fast at 2.5x and Ultrafast at 8x.
Riyal amounts are approximate at the official rate of SAR 3.75 to the dollar.
Frequently asked questions
When do ChatGPT usage limits change?
In December 2026, according to OpenAI's help page. No exact day has been announced. On Plus, Pro and Business, usage becomes one allowance shared by ChatGPT and Codex and measured in tokens.
Will ChatGPT subscription prices go up with this change?
OpenAI has announced no price change. Plus stays at $20 a month, Business at $20 per user billed annually or $25 billed monthly, and Pro at $100, $200 or $500.
Does regular ChatGPT chat count against the new allowance?
No. OpenAI states that Instant chat will not draw from your usage. Advanced models, long tasks and Codex work do, and Fast mode uses the allowance at 2.5 times the normal rate.
What happens when an employee runs out of ChatGPT usage?
They can wait for the reset, upgrade the plan, or buy credits to keep using eligible features. Remaining usage and reset times appear in Settings, then Usage.
