Codex eating your quota? A 1M-token context setting is killing your cache

altryne · x · 2026-09-27

The author traced sudden Codex quota drain to an old experimental setting: a 1M-token context window, which destroys OpenAI's cache and rips through limits. After removing it and moving sub-agents to a cheaper model, usage went back to normal. Lesson: check your codex.toml for leftover experimental settings.

Related event: Codex users warned: 1M context setting destroys cache and burns quota(3 posts)→

Original post →

More from coding & agent

coding & agent channel →