OpenAI will combine ChatGPT web chat and Codex usage limits in December
Prepared by Sintelligent with the help of SI. Published in SI: . Updated: .
Coding and ordinary conversations will become parts of one computing budget: active users will have to make more conscious choices about which tasks to spend it on.
OpenAI told Plus, Pro and Business subscribers that starting in December, the ChatGPT web chat and the Codex coding agent will draw on one shared limit instead of two independent ones. Quick responses in Instant mode will not consume this resource. Metering will shift from the number of messages to tokens processed and generated, so usage will depend on the model and task complexity. The shared balance and reset time will appear in usage settings; after exhaustion, users will be able to wait for a reset, change plans or buy additional credits. OpenAI links the change to moving ChatGPT Work's agent capabilities into the main interface and promises that most users will not have to wait for limits to reset more often than they do now.
Sources 1