OpenAI's 5-Hour Codex Limit Drains in 24 Minutes, Frustrating Plus Users

OpenAI recently reinstated a 5-hour usage cap for Codex on its Plus tier, but users are finding the allowance exhausts far faster than the name implies. One developer reported that a single task — adding Find and Replace functionality to a Markdown notes app — consumed the entire 5-hour quota in roughly 24 minutes. The task was left incomplete when the limit ran out mid-verification, leaving the user unable to instruct Codex to finish its own work until the allowance resets. The core complaint is not about having limits, but about the unpredictability of the 5-hour window, which makes it impossible to plan a meaningful coding session. Users argue the cap is especially ill-suited to agentic workflows, which involve codebase exploration, testing, and browser verification — all of which consume quota rapidly and unevenly.