Over the past few days, I’ve noticed something with Codex that has become increasingly frustrating. The weekly limits are starting to feel extremely restrictive.
Back when we had the 5-hour limits, I was actually able to get more work done, and even when I used Codex very aggressively and heavily, my limits would rarely run out in less than five days. Now, the weekly limit seems to drain almost as quickly as the old 5-hour limit did.
My usage hasn’t changed. I’m using Codex the same way I always have:
No MCP or anything similar.
Just plain Codex Desktop.
My workloads are roughly the same.
Yet the limits are being consumed significantly faster.
On top of that, even after Tibo resets my limits, they sometimes run out again in less than a single day.
Is anyone else experiencing the same thing? How many people have noticed their weekly limits draining much faster recently?
I really hope this is some kind of bug, because if this is the intended behavior going forward, I’m seriously considering canceling my plan.
I usually use 5.6 Sol on High or Very High. For smaller tasks, I use Luna or 5.5 High mode. Sometimes I have Fast enabled as well.
I use Ultra very rarely—maybe once or twice for a one-off planning task. This was already how I used it before.
When the weekly limits were first introduced, I could barely use them up with the same workload. In fact, I was working on two projects at the same time back then, and the limits still weren’t dropping to these levels.
I feel every single word you’re saying! During those two resets within such a short space of time, I missed one and saved the ~86% for heavy tasks. Then there was the reset I didn’t even notice. I started my heavy task, assuming everything was fine. Now I’m at 0%, effectively two missed resets, and I have to ‘wait’ for almost a week. I actually use the pro subscription. ‘Banked resets’ would definitely be the smarter choice rather than executing random and completely unexpected resets.You feel forced to constantly squeeze the maximum out of your tokens because you can never be sure when a reset is coming, or if one is coming at all. This makes scaling unpredictable and encourages waste where none is needed. It makes a subscription via OpenAI inefficient and unpredictable for me. Saving tokens makes no sense, wasting tokens doesn’t either. OpenAI urgently needs to figure out which direction they want to take—and then live with the consequences.
It’s not that usage has increased; rather, the weekly usage limit was reduced after the reset. My new 20x account subscription has a weekly usage limit of only $200, but the account I registered a year ago—which is now subscribed to a plan with a weekly usage limit of over $2,000—hasn’t been affected at all. I suspect the reset caused the weekly limit to be reduced.
I’ve had the $200 Codex plan for months, but this is the first time I’ve experienced anything like this. Seeing that others feel the same way is reassuring. Hopefully they hear us and find a solution.
It feels like the 100% limit/usage amount we see after each reset changes. Almost as if there’s a different multiplier applied with every reset.
I don’t think this is normal. Either there’s another issue with the limits, or they’ve actually reduced them. If it’s the latter, I’ll cancel my subscription immediately.
A standard 20x subscription has a weekly usage limit of around $2,000, while the Plus plan has a weekly usage limit of around $100. My new account has a weekly usage limit of only $200 following the most recent reset.
Yes. And while people pray to Tibo for resets, I literally do NOT want any resets. Just give me the MUCH higher usage limits we had and pay for, that are predictable to work to. They have dropped a significant amount on the 20x Pro plan. It’s very frustrating. All of this gamifying reset roulette is horrible for pro users. OFC the “friendly usage resets” are to get people hooked on a higher usage baseline that they get comfortable with, because right on cue, the resets slow down and paid resets are here… /sigh
As I’ve stated elsewhere I believe Max and Ultra simply aren’t worth the expense.
Sol High for coding tasks is very good indeed.
The only criticism I have is Codex isn’t yet capable of doing “everything in one go” imho. You have to tackle things piece by piece. However it’s very good at that.
I’m not convinced you can "fire and forget* long tasks reliably either and that modus operandi will land you with too much rework and wasted quota.
As far as “shrinking quotas” are concerned I’m working on a tool to make that more obvious if it’s happening at all.
I have the 5x plan, and the % usage rate burn sure seems to be a lot higher with not a lot of usage and I have been on 5.5 medium or Terra Medium trying to keep it friendlier.
You’re not alone. Since release 5.6, the situation has been dire. Week after week, new version updates generate new, increasingly serious problems, including performance issues at a tremendous cost.
This is the moment when, out of love for Codex, I begin to feel the urge to betray it.
same issue for me, I’ve been using SOI Ultra (fast) as a $200 pro user since it was released. But it’s just the past two days that I ran out of my weekly quota in a couple of minutes! And I have been doing the same thing that I used to do, Nothing new!
I’m experiencing exactly the same issue. My Pro 20x usage allowance is being consumed several times faster than before, even though my usage pattern hasn’t changed significantly.
I paid $200 for the Pro 20x plan, but right now the actual usable quota feels almost no different from what I would expect from the $20 plan. At this point, the service is practically unusable for the workload I subscribed to Pro for.
If this is a bug, OpenAI needs to acknowledge and fix it as soon as possible. If the issue continues without a proper resolution, I will request a refund. There is simply no point in paying $200 for Pro 20x if the actual usage allowance is nowhere near what the plan is supposed to provide.
Sol Max is definitely worth the expense, specially when you’re working with multilevel logic and pattern discovery. Xhigh was tried and it just doesn’t cut the most sophisticated tasks, specially erecting statistical analytic structures from the ground up.
I’ve been on 5x for over a month now and up until now I barely had to worry about even looking at my usage. I mostly use codex after work and on weekends. Recently, on a day off, I used up over 50% total usage in just over a day.
I’ve now got about 5 days to go to reset with less than 30% left. That’s way too fast and never experienced before.
Also, I mostly only use Sol on xhigh or high depending on what I’m doing.