IT Home On October 16, according to a report, @thsottiaux, the person in charge of Codex / ChatGPT at OpenAI, posted a tweet on the X platform yesterday (October 15) stating that the team has implemented a series of changes and increased computing power, fixing the error “the selected model has reached capacity limit, please try other models.”
IT Home note: The OpenAI team is currently advancing the “28 Days Continuous Delivery” program. In the 11th announcement, it was announced that a series of improvements and additional computing resources have been added, with the goal of significantly reducing the frequency of this error.
Regarding issue feedback, based on public discussions, this problem is not simply that “the model is actually fully loaded”. Some users reported that even when the subscription quota is sufficient and different models are switched, the error still occurs repeatedly; some developers suspect it is related to backend requests, session state, or resource scheduling.
In terms of deployment measures, the OpenAI team optimized the system and invested more computing power to reduce model capacity errors. Secondly, when errors still occur, tasks should be able to “continue automatically in the end”, rather than staying stuck in an error state and waiting for user intervention. The team also invited users who are still encountering problems to provide feedback so that further troubleshooting can be carried out.