OpenAI plans shared usage limit for ChatGPT and Codex in December
Instant conversations will remain outside the limit, while allowance sizes and reset schedules have not been disclosed.
Subscribers to OpenAI's Plus, Pro and Business plans will have a combined allowance covering ChatGPT and Codex beginning in December, the company says. Its Help Center outlined the arrangement in a November 2 update but did not identify an exact launch date.
The system will measure consumption in tokens—units representing text that models process or produce. According to OpenAI, more sophisticated models and complicated tasks generally consume more tokens. Codex coding sessions will leave less capacity for demanding ChatGPT requests, and those requests will likewise reduce the allowance available for Codex. Instant conversations, however, will be exempt.
OpenAI linked the decision to bringing Work's agent capabilities into ChatGPT. The company expects most users will face no increase in how often they must wait for an allowance reset.
The update provides neither plan-specific allowance sizes nor a reset schedule, and it makes no mention of Free or Go subscriptions. ChatGPT's Usage settings will display the remaining balance alongside reset times. After exhausting the allowance, subscribers can await a reset, move to a higher plan or purchase credits for features that support them.
