Tell HN: OpenAI brings back 5 hour limit for plus and business standard users

2 points by spwa4 ↗ HN
In case you're wondering why the limits behave so very different from last week. Also: this makes limit resets kind woth significantly less.

68 comments

[ 0.33 ms ] story [ 43.5 ms ] thread
In hindsight, this makes more sense with the release of Astra. The $100/mo is now a more compelling upsell.
I went another way and signed up for a opencode go subscription on the side.
Opencode Go's value is pretty poor these days. It was better a month or two ago.
Session limits are obnoxious and make Codex much less useful for me. I tend to code in spurts when I find some time and the $20 weekly limit was reasonable for my side projects.

I'm now hitting the session limits which means I can either purchase another plan, upgrade, or move off of Codex.

I combine it with OpenCode + a small OpenRouter budget. GlM 5.3 Flash, Luna, Gemini 3.7 Flash and a couple of other very cheap models are sufficient for a lot of tasks if put on the right road. Telling Astra to debug an issue should be a last resort, it would blow through 10% of the session limit, but cost like 10c on one of the open models. Same goes for exploration, documentation, configuration and smaller features. These models are generally good enough, and you can still do a review with a strong model for a fraction of the cost.
This allows users to pace their subscription usage to last the whole week.

I've had few weekends where I spend the full week credits over night then have nothing to do for a week

It might be good to have at least two hobbies.
This is very fair. Compared to Claude the Codex limits are much better and smarter model.
Clearly, the prices and limitations will increase with time if they don't find an efficient way to perform training and inference of LLMs. It would be great to see those issues being seriously addressed and, eventually, being fixed for good. THAT would definitely make an important and practical difference. Not only economically but also scientifically.
I always wonder where the people are who continually claim that this whole thing is already profitable are when prices are increased or services cut.
I’ve not heard anyone say that the whole thing is profitable. Companies like Anthropic have said that inference is profitable, but I assume that’s only on a steady state basis, which nobody in the industry has ever reached at this point. As far as I know, everyone is still burning cash with data center buildouts and overall training costs.
I think the people that are saying that mean that the inference is massively profitable. From what I've seen, the margins are between 50 and 90% on an API call. It's the training that's very, very expensive!

As far as I understand, for both OpenAI and Anthropic, currently the majority of their GPUs are being used for training, with only a smaller portion being used for actual customer serving.

Inference requires training. That's a disingenuous point.

It's similar to people talking about drug companies charging hundreds of dollars for a medication that costs dollars to manufacture. Nevermind that it took them 15 years of research and trials to discover it and get it through trials and the approval process. Nevermind again all the drugs that failed at one of those steps.

I dislike this approach. In my opinion it is much useful a 1x/2x billing based on hour like Deepseek does. Being unable to use it at the hours I need it makes me want to remove the service, not upgrading it.
This is one of the reasons I’ve built a resident daemon on top a local CC/codex. 5h sessions is a [lazy] way to shift the scaling responsibility onto a user, so they are staying with us for a while
People should realize by now that they aren't giving out thousands of dollars worth of compute in a $20 subscription out of the goodness of their hearts. The base level plans aren't meant for any kind of serious work beyond the equivalent of Google searches. At best treat it like a trial.
I hope everyone recognized the cheap tokens as a transparent ploy to gain more users before jacking up prices. This was obvious from day one.
[delayed]
It’s good to be aware when you’re buying something that’s subsidized. No need to downplay the importance of awareness.
agree and underscores the idea that you should keep your setups portable between model providers by using external memory systems and connector gateways
Only if switching costs are low. If you are investing in the platform and integrating deeply, then planning to switch may not be a good plan.
Or to just give users enough usage to find categories in which it is useful for them.

I pay the $200/mo., and don’t regret it for an instant - I have a project manager, an executive assistant, a business analyst, a software developer and an international accountant, for an absolute song.

you’re using an LLM as your international accountant? And brag about it online? That’s definitely a choice
It was obvious indeed, but damn does it feel good to have ridden the May 25 - Aug. 26 gravy train.

Now, let's see how the Anthropic IPO goes.

I'm going to predict: soft, below first-trade price at 180d from IPO date.
Is it not due to the all the user data they are gathering?
I thought it was always about us being the RL. We pay less because they use our usage to train their models.
I can't believe people are letting themselves be taken for a ride so shortly after seeing what happened with streaming subscriptions.
Disagree, not for serious work if you're a developer aka requiring loads of iterations tool calls etc. But for other tasks it's relatively fine.
> People should realize by now that they aren't giving out thousands of dollars worth of compute in a $20 subscription out of the goodness of their hearts. The base level plans aren't meant for any kind of serious work beyond the equivalent of Google searches. At best treat it like a trial.

I've said this countless times already; if OpenAI or Anthropic were serious about world domination they would offer an infinite $500 to $1000/month tier subscription. No limits; eat as much as you'd like buffet. Maybe limit concurrent connection (say 10 conns max in parallel) to limit abuse. This would actually allow regular users to run 24/7 agentic loops, massively speeding up deployment. Win win for everyone.

The business model works by over provisioning, which means you actually do want to cap the people who will attempt fully utilize the service.

Once you know someone is going to use 100% of the service rather than 25%, you have to charge them full price. That's why you see them fail over to API pricing.

The $100 plans are plenty for me to use nearly all day. I'm using between $1500-3000 worth of tokens a month via those plans.

I definitely see the pricing model. I can afford $100, maybe $200, but not $3000.

one word, localmodels
They must have realised by now what they are selling is addictive. It's pretty normal for drug dealers and the like to give people introductory prices then hike once dependence kicks in.
> They must have realised by now what they are selling is addictive. It's pretty normal for drug dealers and the like to give people introductory prices then hike once dependence kicks in.

I have never heard of a drug dealer actually doing that.

How is 5 hrs being measured and over what time period?
it is a window that starts with usage and continues for five hours.

When I was doing an evaluation using a lower paid tier of Claude, I would have a service send a "hello" ping 4 hours before I started my work day, to reduce my first work-hours session window to 1 hour.

The goal was to have this be more of a "thinking" session for planning the next larger block of work, and then being able to use a lower cost model for implementation.

That said, if your 5 hour window quota is 15% of your weekly quota, this means you can be using 30% or more a day of the weekly quota.

Once they capature enough marketshare prices and restrictions are both going to skyrocket. The difference between the subscription usage and API pricing are stark.
Where's the moat? You can only truly "capture market share" if there's lock-in and no competition.
Harness and data seem like two vectors. It's also not clear to me that at some point we won't get a snowball effect from the platform that has the most use and thus the most training data just running away with a positive feedback loop.
Controversial take but this works better for me than just the weekly limit. I have about 1-2 hours of sustained focus per 5 hour period so if I run out of tokens then I can take a break and start again in a few hours. Or just use the top ups that I’ve paid for in case I want to push through.

Having just the weekly limit meant I could blow through too much within the first couple of days. Fortunately, resets were raining down during that time.

(I know I could just vibe code a tracker that splits up my usage into arbitrary periods and lets me know when to take a break. Maybe if they remove the 5-hour limit again.)

What also sucks is your weekly reset does not also reset your 5 hourly
Oh that's just sloppy of them.
I think it's reasonable for the Plus plans to have this. If you're doing any serious work you should be on the Pro plan. If you're a casual user, 5 hours is more than plenty.
5h limits are awful. It means it is literally unable to complete a large task.

A better approach if they want to balance the load is having higher usage or lower usage consumption at different times of day.

OpenAI must be under a lot of pressure to make the books balanced. A bleak sign for Anthropic's IPO and the stock market.
My goodness, the complaining... Just get all your devs a $200 ChatGPT Pro (20x) plan. Yes, you lose the "team" component, but you gain so much more. And what's $200 against the salary of a good developer? It's absolutely inconsequential, even in far cheaper non-US salary regimes.
(comment deleted)
Sam, is that you?
I think Sam would be saying something pithier. And the 20x Pro plan absolutely runs at a loss, so I think he would be the last one to promote it.
Think they are pretty happy with 20x pro subscribers, even at a loss. Highest value customer base, they'll be happy to take a loss to gain market share.
Devs are already on an enterprise plan. I think the people complaining are the personal use folks
i would love to give them $200 a month, but unfortunately I have no money

AI is just another case of the rich getting richer, people who dont really need AI with practically unlimited access, and people who need to it to try and change their lives being priced out

its a wealth inequality accelerator at the worst time in history for wealth inequality

i would take 50% less usage just to be able to use it in my own time, the plus plan without 5 hour limit gives me 1 days work a week, so 4 days of usage per month, and I was happy with that

now thats been taken away i get 15 minutes of usage then 5 hours later ive lost interest in the work, i now sit with 70% usage knowing 5 hours ago that a reset was coming and I couldnt even be bothered using it

its completely killed the product for me, im now paying 20 a month for something that is no use to me

At least we all got a reset out of it! I was holding off, refreshing & smashing "beg" on https://codex-resets.com

I appreciate that he gave the heads-up. I deliberately burned through my quota yesterday, low key expecting a reset, but ready if I needed to use a other banked reset.

> Thanks for reading. We will do a global reset of the usage for all paid subscriptions so that you can keep enjoying Astra after burning through all of it doing fun 3D modeling in blender. The work week is about to start.

> Lands around 6pm PST today.

https://m.x.com/thsottiaux/status/2097043464538264003

Haven’t these been back for weeks now?