Quota depletion for some programmers using the Codex tool accelerated this week due to technical inefficiencies in the system's caching process. Tibo Sottiaux, representing the service, stated that lower cache hit rates compared to previous stable periods caused usage limits to drain more rapidly for affected users. OpenAI is investigating the root cause of the decline and plans to provide a status update tomorrow.
The acknowledgment follows claims from thousands of users regarding rate limit reductions, which OpenAI initially attributed to the use of unsupported tools such as sub2api. This week, the company also announced that Codex reached 20M active users and issued banked usage resets to all Codex and ChatGPT Work subscribers to mark the milestone.
Key sources
- SOURCE@thsottiaux“We are investigating and will have an update tomorrow.”x.com