Make the temporary removal of the 5-hour usage limit permanent
Feature request
Please make the current temporary removal of Codex's 5-hour usage restriction permanent for Plus, Pro, and Business plans, while retaining the existing weekly usage allowance.
Context
On July 12, 2026, Tibo Sottiaux announced that OpenAI was temporarily removing the 5-hour usage-limit restriction while rolling out GPT-5.6 efficiency improvements:
https://x.com/thsottiaux/status/2076365965915467978
After using Codex under this weekly-only system, I believe it is a substantially better way to manage included usage and one of the most meaningful improvements to the product since the release of GPT-5.6.
Why the weekly-only model is better
The 5-hour window requires users to continually monitor a short-term meter and organize their work around rolling reset periods. That creates friction precisely when Codex is most valuable: long-running development, debugging, refactoring, research, and other tasks whose natural duration is difficult to predict.
With a weekly allowance as the governing limit:
- Users can focus on completing work instead of micromanaging consumption within the 5-hour usage window.
- Longer and more ambitious tasks are less likely to be interrupted at arbitrary points.
- Users can distribute their included usage according to their own schedules and workloads.
- The overall weekly allowance still provides a clear, finite usage budget.
- Codex feels more predictable and suitable for serious project work.
This is not a request for unlimited usage. It is a request to let customers decide how to allocate the usage already included in their weekly allowance, without an additional short-window restriction determining when that allowance may be used.
Requested behavior
- Permanently retire the 5-hour usage window for Plus, Pro, and Business plans.
- Retain the weekly included-usage allowance.
- Continue applying reasonable fair-use, abuse-prevention, and infrastructure protections where necessary.
- Clearly document the weekly allowance and how usage is calculated.
- If capacity protection still requires a shorter safeguard during exceptional demand, use a transparent temporary mechanism rather than restoring the 5-hour window as the normal customer experience.
Why this experiment matters
The temporary removal has demonstrated an important product benefit that usage statistics alone may not capture: reduced cognitive overhead. Not having to repeatedly check a meter makes Codex feel calmer, more trustworthy, and easier to integrate into real work.
Please evaluate the experiment not only by aggregate consumption and infrastructure load, but also by task completion, interruption rates, sustained project workflows, and user satisfaction. The weekly-only experience is a major improvement. Please make it permanent.
19 Comments
+1
It would be awesome and less mentally stressful 🔥
Bump Up +1
+1 Really good idea
+1 Really good idea
Good idea.
I completely agree. The absence of the 5-hour limit allows for more accurate planning and eliminates the anxiety that a task might not be properly completed because I'm approaching that threshold.
In the end of the day, we are all adults capable of managing our own time without having to adapt to artificial five-hour restrictions.
please
The 5 hour limit is there for a reason. I propose another idea: keep the 5 hour limit, but when 5 hour limit or weekly limit is gone then don't stop all usage just heavily throttle it. Maybe a tiered system like 200$ subs gets to be throttled to 10% speed, 100$ to 5% and 20$ to 2.5%.
This way tasks would not get cut in half, we could start a task at 10% left from the 5 hour window and it would either finish before the window resets or speed up when it resets.
I think this would be very generous from them while balancing the load on their servers better. They could even implement dynamic throttleing like: now the load is high on the server so you get a high throttle, but maybe a bit later when the load is lower it speeds it up for some people to finish their task. This model could give a lot of value for a lot of people while still having some load balancing on the servers.
Absolutely love this idea, the 5 hour window sucks and makes me work on your time, not mine. If I want to burn it all in a day because I'm going to be doing something else all week I can.
+1 feels so much better
@claude, consider this?
Sincerely,
Former 5x, now 4x ClaudeMax20x account holder
Former 1x CodexMax5x, now 1x CodexMax20x account holder
Considering going 3x ClaudeMax20x & 2x CodexMax20x
PS. @codex && @claude consider 100x plans please to reduce Xx statements needed in gh comments && dozens of other obvious efficiency/productivity reasons. :)
Yes please I'm begging: do not put back the 5h limit.
My first day of using 5.6 Sol i started with a big task and it went on to explode the 5h limit and spend all my stored credits (~40€) and then in addition spend an extra 45€ to recharge credits because of that 1 prompt while the weekly limit had still some margin. I'm on the Plus plan so for me that one task cost me 4 months of subscription just because of the 5h limit.
At this point it has just become a barrier to do big tasks, while this is not a problem with the weekly limit only. So please do not put it back it will just reduce de significance of the work we're allowed to do.
This is my main gripe with Codex now. I can't stand the 5 hour limit and the mental block it puts on my creativity
The 5-hour limit is making my work genuinely miserable. It completely destroys any kind of deep, uninterrupted workflow - especially when working on larger coding tasks.
Instead of focusing on the problem, I constantly have to watch the usage meter and worry that one heavy task will burn through a huge part of the limit. That makes the whole experience stressful and frustrating.
A weekly limit is understandable. But the 5-hour limit makes serious daily work feel unnecessarily painful. Please reconsider it.
The current 5-hour Codex limit is too restrictive for serious development work, especially for long-running agentic tasks.
In practice, I regularly hit the 5-hour limit while still having a large majority of my weekly allowance remaining. This means I cannot actually make effective use of the weekly quota I am given. A single feature implementation or larger code-analysis task can already span multiple 5-hour windows.
The combination of a weekly limit and a restrictive 5-hour limit effectively turns the 5-hour limit into the dominant constraint. The weekly allowance therefore becomes misleading: even if substantial weekly capacity is still available, I cannot use it when I need it.
For software development, usage is naturally bursty. There are periods where Codex is used intensively for several hours during implementation, debugging, testing, and review, followed by periods of little or no usage. A short rolling limit works against this workflow.
I would strongly prefer one of the following:
The important point is that users should be able to decide how to distribute their weekly Codex allowance. If I still have most of my weekly quota available, I should not repeatedly be prevented from completing a feature because an additional short-term limit has been exhausted.
O retorno do limite de 5 horas no Codex é muito restritivo. Antes, com o limite semanal, eu podia distribuir meu uso conforme minha necessidade e até concentrar grande parte da cota em um único dia. Agora posso ter quase toda a minha cota semanal disponível e ainda assim ser impedido de usar o Codex por várias horas. Seria muito melhor manter apenas o limite semanal, ou permitir que o usuário consuma livremente sua cota semanal.
We want to remove the 5 hours limit again open ai
I am not asking for more quota. I am asking for more control over when I spend the quota I already have.
+1