Overview
- OpenAI announced that paid Plus, Pro and Business subscribers will draw ChatGPT and Codex usage from a single token-based quota beginning in December.
- Usage for both products will be measured in tokens and actual token burn will vary by model choice and task complexity, so heavier models will consume quota faster.
- Instant Chat will remain exempt from the shared quota, and users will see their merged remaining balance and next reset time in Settings → Usage.
- If a user exhausts the shared quota they can wait for the automatic reset, upgrade their subscription, or buy additional credits to continue working.
- OpenAI says it is rolling out the change to avoid workflow disruption and expects most users will not need to wait for resets more often than before, though heavy code or multiagent work could change how quickly individuals hit limits.