Paid members of the ChatGPT platform will see changes to how their quotas are managed when a new metering policy takes effect across its tool suite. Beginning in December, a unified allowance will apply to the Plus, Pro, and Business plans, combining the usage for both the ChatGPT chatbot and the Codex coding tool. This new system will be based on tokens rather than message counts, meaning that the use of more advanced models or the submission of complex tasks will deplete the shared allowance more quickly, though requests made via the “Instant” chat feature will not be counted.

The confirmation follows an Oct 29 update in which OpenAI doubled capacity for Business Premium and its Pro 100, 200, and 500 subscriptions, increasing the top tier’s allowance to 50 times that of the $20 baseline plan. The company also canceled a proposed reduction in limits for Pro 200 subscribers and launched a development sprint focused on application speed and simplicity. For the time being, extended limits for the company’s “dots” feature will continue to be tracked separately from general usage quotas.