# ChatGPT and Codex Shared Usage Limits: What Changes in December

OpenAI plans a combined allowance for Plus, Pro and Business, but the missing quota details matter more than the new meter.

By Paras Madan | 2026-11-01


OpenAI says ChatGPT and Codex will share a usage allowance on Plus, Pro and Business plans starting in December. The announced change brings Work's agentic capabilities into the main ChatGPT experience. **Instant chat will not consume that allowance**, an important exception to a shared-limit headline.

This is an announced transition, not a change that subscribers should assume is already active. The [new usage-limit documentation](https://help.openai.com/en/articles/20001625-understanding-chatgpt-and-codex-usage-limits) gives the month but no exact start date.

## How the shared ChatGPT and Codex allowance will work

OpenAI will measure usage in tokens, with consumption depending on the selected model and task. More demanding work generally uses more allowance. Subscribers will see their combined remaining usage and reset times in Settings → Usage.

After reaching a limit, the documented options are waiting for a reset, upgrading, or buying credits for eligible features. The announcement does not publish plan-by-plan token quantities, new reset intervals or credit prices. It also does not specify a corresponding change for Free, Go, Enterprise or Edu.

OpenAI expects most users will not have to wait for resets more often. That is the company's expectation, not evidence that every existing workflow will retain the same capacity.

## Why this differs from the October Pro allowance increase

The [October 29 release notes](https://help.openai.com/en/articles/6825453-chatgpt-release-notes) increased Work and Codex allowances for Pro plans and canceled the planned Pro 200 reduction. Pro 100 and Pro 500 received twice their previous allowance; lower-allowance Pro 200 subscribers were brought up to the higher level.

December's announcement concerns which activities draw from an allowance. Increasing a balance and changing what spends it are different decisions. One does not tell a subscriber the practical effect of the other.

For example, someone who uses ChatGPT for long analytical tasks and Codex for repository work should evaluate those activities together after the transition. A larger nominal allowance can still feel different when two substantial workloads share it. Someone mainly using Instant chat may have a different experience because that activity is explicitly excluded.

## What subscribers still need to know

The missing quantities prevent a reliable before-and-after comparison. A token-based meter alone does not establish a price increase, a capacity cut or a better-value plan. Those conclusions need the actual allowance and representative consumption under the new rules.

The [Hacker News discussion](https://news.ycombinator.com/item?id=50320588) shows why the distinction matters: some commenters value keeping coding and demanding ChatGPT work separate. Those comments describe individual preferences and concerns, not measured outcomes from a transition that has yet to happen.

The useful comparison in December will be completed work between resets, with similar models and settings. Until OpenAI publishes the missing numbers, **a combined usage display should not be mistaken for a complete explanation of subscription value**.

I'm Paras. I work at the intersection of AI, engineering and distribution.

I help AI founders explain product changes and technical launches in content that developers and founders want to read and share.

[Book a 30-minute call](https://cal.com/paras-madan/30min).