OpenAI to merge ChatGPT and Codex usage into one shared quota from December

OpenAI announced via a help-center document that starting in December it will combine ChatGPT and Codex usage into a single shared allowance. Instead of billing per message, usage will be measured in tokens. The pool applies to Plus, Pro, and Business subscribers, whose consumption across chat, Agent features, and coding sessions will be drawn from the same bucket.

According to the report, ChatGPT chat will also start drawing down allowance going forward. When users handle more demanding tasks inside a chat, such as generating images or drafting content, those activities are billed according to usage rather than being free. The one feature that remains outside the billable scope is Instant chat, which continues to be free of charge. Notably, this is not the same as ChatGPT Work: the Work allowance had already been merged with Codex, and the current change brings ChatGPT chat itself into the counted usage.

The announcement arrived through a help-center page rather than a launch event or major press push. Both reports were published the same day, and at the time of writing neither contained further detail on the exact rollout date or on how usage would be measured in edge cases. For users who split their time between conversational ChatGPT and Codex coding sessions, the key shift is that the two usage streams will no longer sit in separate buckets once the change takes effect in December.