OpenAI to unify ChatGPT and Codex usage limits in December 2026

The move to a shared pool

OpenAI has announced a significant shift in how it manages subscription limits for its primary AI tools. Starting in December 2026, the company will introduce a shared usage allowance that combines ChatGPT and Codex for several paid tiers. This means that the resources used for advanced conversations and software development tasks will be drawn from the same bucket.

This change specifically targets Plus, Pro, and Business subscribers. While these users will see their limits unified, the company has stated that Free, Go, Enterprise, and Edu accounts will not be subject to this specific December change. This creates a distinct experience for paid users who rely on high-level capabilities for professional work.

The transition is part of a broader effort to integrate various tools into a more cohesive experience. For example, OpenAI is working to bring the agentic capabilities of ChatGPT Work into the main ChatGPT interface. By unifying the limits, the company is essentially merging the 'fuel' available for different types of complex, multi-step tasks.

How tokens determine your limits

The new system will measure usage based on tokens, which are the fundamental units of information that AI models process and generate. Because different tasks require different levels of computational power, the amount of allowance consumed will vary depending on the complexity of the request. More demanding workloads will naturally use more of the shared pool than simpler queries.

For instance, a developer who uses Codex to perform a deep code review and then switches to ChatGPT for a complex technical analysis will now be drawing from the same total allowance. This could change how teams prioritize their tasks throughout the day. If a user exhausts their limit on heavy coding tasks, they may have less remaining for conversational analysis.

OpenAI has not yet released the specific numerical limits or the schedules for when these limits reset. However, the company expects that most subscribers will not find themselves waiting for resets more often than they do under the current system. Users will eventually be able to see their remaining combined allowance and reset times directly within their account settings.

Impact on developers and teams

This policy change creates a new landscape for organizations that rely on both ChatGPT and Codex for their daily operations. IT teams and developers will need to find new ways to calculate and forecast their future consumption. Because the accounting rules for individual features are not fully detailed, it is currently difficult for companies to compare the new limits against their existing subscription costs.

The move comes shortly after another update on October 29, 2026, which doubled the allowances for Pro 100 and Pro 500 subscribers. Those on the higher-priced plans may need to perform a new cost-benefit analysis to see if the shared pool provides enough value for their specific needs. Subscribers comparing different Pro tiers will need to wait for official figures to make an informed decision.

Teams that are currently comparing different coding agents, such as Codex and Claude Code, may also need to rethink their development workflows. They will have to consider how a shared limit affects the speed and volume of their software development activities. The lack of transparency regarding specific numbers means that many organizations will have to wait until December to see the full impact.

What stays the same

While the shared allowance is a major shift for advanced features, some parts of the ChatGPT experience will remain unchanged. Specifically, instant chats will be exempt from the new shared usage accounting. This means that basic, quick interactions will not count against the pool reserved for more complex tasks like coding or long-form reasoning.

Codex Cloud also remains a unique entity in this ecosystem. It is restricted to eligible paid plans and continues to be subject to specific rollout and workspace settings. This distinction suggests that while the core tools are being unified for many users, specialized cloud environments may still operate under different rules.

For users who do hit their limits under the new system, OpenAI will provide several options to continue working. Subscribers who reach their cap can choose to wait for the next reset, upgrade to a higher tier, or purchase additional credits for eligible features. The goal is to provide flexibility while moving toward a more integrated model of AI consumption.

Preparing for the transition

Because OpenAI has not yet published the specific numerical limits for the December 2026 change, organizations have some time to prepare. The best course of action for businesses is to establish a baseline of their current usage across both ChatGPT and Codex. By tracking how many tasks they currently perform in each tool, they can better estimate how the shared pool will handle their workload.

As the deadline approaches, it will be important to watch for the official publication of the allowances for each plan. These figures will be the primary data point for deciding whether current subscription levels are sufficient. If a company finds that their heavy coding needs are draining the pool too quickly, they may need to adjust their plan selection before the policy goes live.

The shift toward a shared allowance reflects a trend toward