OpenAI to merge ChatGPT and Codex token allowances in December 2026

The move to a shared pool

OpenAI has announced that it will combine the usage limits for ChatGPT and Codex into a single allowance for its paid subscribers. This change is scheduled to take effect in December 2026 for users on the Plus, Pro, and Business plans. The integration means that instead of having separate buckets for different tasks, users will draw from one shared credit balance.

The new system will measure usage based on credits per million tokens. Tokens are the small units of text that AI models process to generate responses. By unifying these pools, OpenAI is bringing its coding agent, Codex, more fully into the existing ChatGPT ecosystem. This shift represents a significant change in how paying customers manage their daily AI interactions.

As reported by Crypto Briefing, this move is part of a broader effort to streamline how OpenAI handles its various products. While the allowance becomes shared, the core plan structures will still differ. For example, Pro plans do not carry a 5-hour limit, whereas Plus and Business plans do. Pro users may also continue to face specific weekly caps despite the merged pool.

Transitioning to token-based pricing

This merger follows a major pricing overhaul that OpenAI initiated on April 2, 2026. During that time, the company shifted to a token-based usage model measured in credits per million tokens. Before this transition, users were billed based on a per-message estimate, which often failed to distinguish between different types of requests.

The old per-message system treated a simple question and a complex, compute-heavy request as roughly the same unit of work. The current token system was designed to more accurately reflect the actual computational demands of various tasks. This is particularly important for advanced models like GPT-6, which require more resources to process complex instructions or large amounts of data.

OpenAI has been refining these limits and pricing structures over several months. For instance, the company doubled the allowances for Pro 100 and Pro 500 in October 2026 and canceled a planned reduction for Pro 200. These adjustments show a pattern of tuning the balance between user access and the high costs of running large-scale AI models.

Impact on developers and businesses

For developers who frequently switch between writing code and chatting with the AI, the shared pool removes a specific point of friction. Previously, a user might run out of capacity for Codex while still having plenty of credits left for ChatGPT. The new unified balance ensures that credits can be used flexibly across both tools regardless of the specific task at hand.

However, this shared pool introduces a new form of internal competition within a single account. A team that relies heavily on Codex for software engineering might find that their general ChatGPT usage is squeezed. Because every prompt and code generation will draw from the same well, users will need to be more strategic about how they allocate their daily limits.

Business customers may find the new framework more favorable for long-term planning and financial forecasting. Having a single metered balance is often easier for companies to track than managing two separate sets of limits. The research suggests that these more predictable and user-friendly frameworks could make it easier for organizations to adopt these tools at scale.

What to watch for next

While the merger is set for December 2026, several questions remain regarding how specific limits will interact with the new shared pool. It is currently unclear how the 5-hour limits on Plus and Business plans will behave once Codex begins drawing from that same balance. Users on these plans may find the restrictions more noticeable as their coding tasks begin to impact their chat availability.

The history of OpenAI's pricing suggests that these numbers are subject to change. The company has adjusted allowances multiple times, including temporary doublings and canceled reductions. Anyone building specific workflows around current limits should remain aware that the exact numbers could shift again as the company refines its strategy.

Users who split their time between chatting and coding should begin monitoring which activities consume the most tokens now. By tracking current usage patterns, subscribers can better prepare for the transition. Every prompt and every line of code generated will eventually count toward the same total allowance in the coming year.