OpenAI Codex Update: 94% Faster Loading, 98% Fewer Requests

OpenAI showed the results of a major Codex performance update that is set to be released next week.

It should noticeably speed up the loading of very long conversations while reducing memory consumption and the number of requests.

In an extreme test with a conversation of 741 messages and a volume of 231 MB:

• loading time dropped from 27.6 to 1.7 seconds
• the number of requests decreased by 98%
• the number of loaded conversation items became 99.6% smaller
• the app's memory usage growth decreased by 41%