Hacker News

All these resets are given because of rival competition. When they win the market, we are going to be squeezed $$$.

Quietly eh?

Yeah, it's the gacha of software development. I too redline my subscription every single week just because I want to get the most value out of the money I spent. Literally 100% weekly usage, every single week.

I am using the resulting software and documents though. I'm daily driving the projects I wrote with AI. Started new ones for fun, such as decompilation projects for my cherished childhood games. Hardened my router and my LAN. I just keep finding an unending amount of things to do.

It took over a month to complete a full Fable code review of my lone lisp project.

That's what half of people's AI developed projects feels like to me. I have colleagues using Openclaw, and having an entire team of agents managing their home infrastructure. And I'm just thinking "why?". They're not developing anything that useful, their homes are already automated via home assistant. It's just agents for agents sake. Outside of my work life, I find very little need to be using AI.

AI in general is a slot machine and it wouldn't surprise me if that was a large part of the psychosis. You prompt the lever, watch the display spin, and sometimes, just sometimes, a bunch of coins come out.

We’re at the gamification stage of motivating people to use your product, eh?

Permafable got announced earlier this week, but definitely the slot machine feel is real.

Why does Fable usage end today?!

Anthropic won't have anything competitive with OpenAI Sol and Kimi K3?

Some of these resets were because of bugs in the 5.6 launch that caused subagent-users to burn _a lot_ of extra tokens. It was placating customers.

It is impressive they've got the capacity to do so.

They can do it often because they have compute, and they have compute mostly because they are pretty far behind Anthropic when it comes to heavy enterprise users. They’ve supposedly added several million since 5.6 launch, which is a steep growth curve. We will see if those users stick around. Anecdotally, I’ve stopped using it for anything major because it has attempted to do some very unsafe things when I wasn’t looking. Friends I’ve talked to have also gotten over the initial honeymoon period.

tbf they turned previous ChatGPT macOS app into Codex, so getting a "new" Codex user might mean someone using ChatGPT just updated their app.

I have an inkling of a feeling that this generosity is a stunt only to help juice the growth and “story” specifically for the IPO…

Best use the AI subsidies while you can!

I was thinking about this. How do they measure this? Every upgrade of the ChatGPT apps get counted?

They, like every startup ever, count users in a very stupid way.

One person using the app on two computers and a phone? That's obviously 3 users.

Growth Hackers and SEO types have done so much to make the world an insufferable place. I congradulate them on the OpenClaw psyop, that was some very inspired bullshit that did fuck all and is now irrelevant.

They just merged the ChatGPT app and the Codex app into one super app.

It is a clever growth hack.

Ant is not giving lot of gifts though :(

API pricing itself might have extreme margins compared to the real cost. Anyway, in my testing the Pro 20x $200 plan gives you about $2200 API-equivalent weekly usage if you're only using GPT 5.6 Sol, so quite close to $9k-$10k/month API-equivalent, it's a bit inconsistent with cache costs.

It's very interesting that for Anthropic the $100 and $200 plans only differ 2x in weekly limits, the 5 hour limit differences are more severe. But for OpenAI, Pro 20x is, well, 4x of Pro 5x for only 2x cost. So, for example, 100% of weekly usage for Codex on a Plus ($20) account is just 5% of weekly usage for Codex on Pro 20x.

And you can calculate how much extra usage you can get from resets, and especially banked resets by purposefully using the whole quota and using your banked reset - they expire 30 days after they're given out, so if you don't use one, it just disappears.

How do you spend so much? Our company also has Claude enterprise pricing. I have a monthly limit of $800, I use Opus for everything, and I'm around the limit after the month. Genuinely can't imagine spending 12 times as many tokens without intentionally trying to waste money.

Curious, does Anthropic force companies of a certain size like yours to their Enterprise plans? Or was it by choice for the enterprise features?

Because with a Premium Team seat I run 2-3 vscodes with Opus 4.8 Max all day and never seem to hit my limits.

> Point being my real API usage here must be very unsustainable!

I think it’s funny that everyone anchors to the API pricing as the real cost.

Most likely is that their API costs are printing profits. They can sell the subscription plans at a slight loss because it gets more people like you hooked on GPT models at home and suggesting them at work, where the real money is made.

I think their subscription plans go mostly unused when averaged across all subscribers, too. Some customers are getting great deals by maxing out 100% every week, but most probably use much less.

The reset game is an addictive challenge that gets the hardcore users more hooked on their products because you feel pressured to use it as much as you can before the next unpredictable surprise reset lands.

My usage sky rocketed with GPT 5.6 sol thanks to all these resets. I'm pretty sure this is not sustainable for me.

Kicking myself that I didn't use one of my 4 resets which just expired, but then again, they reset our usage anyway. We are just blasting away with 5.6-Sol-ultra, doing a project I wouldn't have otherwise done (basically drastically editing someone's novel for them and adding illustrations).

Does the company actually gets any value close to your salary + benefits + ~15k$ you gonna burn this month?

Here it's only Microsoft (ghcp) allowed, and 39$ per month is considered ludicrous with no way to increase. Ah Europe

$10,000 in less than three weeks is incredible. The bean counters sure have a blind spot for AI…

yes but as a normal user your data and traces are being used to train AI which helps enterprise customers.

You're using Opus 4.8 and then Sonnet 5 which is both worse and more expensive in practice than Opus.

Price shock in 3...2...1...

> At my $COMPANY, we have claude enterprise at API pricing. I've blown through $10k already this month, opting for Opus 4.8 w/ Sonnet 5 agents where possible.

A very happy gambler of tokens at the Anthropic casino, running up costs on the house at no cost to them, but to $COMPANY paying.

This is the Kimi and GLM effect.

Open Source is coming.

You can use this endpoint for quota reset times.

Just ask Codex to use its local auth token as a bearer token and send a GET request to it. The response includes "available_count" and "credits[].expires_at". Or script it yourself obviously.

https://chatgpt.com/backend-api/wham/rate-limit-reset-credit...

Use CodexBar: https://github.com/steipete/CodexBar

The 5h limit is gone (for now). It does not presently need to be chased. :)

For the automated checking of other stuff: It was a one-shot prompt to get Codex clank up some Python that returns remaining usage, next reset time/date, and so on.

The result does use Codex App Server, but it's a short-lived process that is dealt with over stdio so that's... fine-ish, I guess?

I keep this bookmarked and check it frequently

https://chatgpt.com/codex/cloud/settings/analytics

Check this site for the first part? Maybe they can offer a notifications API, hah.

Oh, they have free drops too? I guess dual wielding the two $20 subs is the best value option right now then.

There were a few tweets about it so they weren't super quiet about it. I think you can get it back by setting `model_context_window=YOUR_VALUE` in ~/.codex/config.toml though.

> I'm starting to think It might be a good idea to stay with American models, by virtue of them feeling so threatened they give out more and more free perks.

If you keep staying with them, they'd feel less and less threatened.

You need to use them about half as much as you do a Chinese model so that the usage stats that get collected show them that they are still losing users (and attention) to the Chinese models.

I've had this too on my personal pro plan and also when I run the usage command it says it's resetting tomorrow rather than getting cut off.

I'm trying to squeeze what I can in the meantime but I feel like they said it was getting cut off multiple times and they keep extending and possibly resetting it (havent been looking close enough to know for sure)

It did once. That's all I know of though. Nothing like Codex.

Nope. Anthropic is incredibly greedy and always has been.

Well yeah they realised that

i just need one more, the last one, i swear

Originally yes, everyone's usage limit went back to 100% at the same time regardless of if you were at 0% or 99%. Now they tend to give out "banked" resets where users can choose when to use it for up to a month (except the most recent one, which I think was non-banked)

My next question is: why? Purely marketing tactics?

Careful with those, they expire.

I just had one randomly delete the other day.

No, they don't apply it :/

Yup! Big beautiful burning pile of cash. Though there’s some nuance with subsidizing inference (which is supposed to make them money long term) vs training (which we’re supposed to ignore I guess).

To be fair, they're motivating me to move from the Pro to the Max tier (the $200 one, which has 4X the usage of Pro for 2X the price).

And they can’t stop because they want to IPO soon, and stopping this madness would signal loudly that they are bleeding cash, don’t really have a business model, and might not be worth $2 trillion after all.

We lose money on every sale, but make it up in volume

I didn't knew about these resets up until a few days back... and every time they used to reset these limits, I thought I was going crazy... the usage numbers just didn't match. And the weekly end window just kept pushing ahead by a day or a few... just the right amount of deviation to make you believe you are going crazy.

Why add a middleman when you can ask your own account with /usage?

Are Codex users even a drop in the bucket of overall use? I mean, it's a pretty specific way to interact with the bot.

Codex usage is clearly common enough to have entered the vernacular of folks here on HN.

But we aren't everyone, and it seems likely to me that there's a lot more people in the world burning tokens using ChatGPT than there are who even know what Codex is.

> can't sustain these resets

If I would have my tin foil hat on I'd say they can manipulate how the subscription usage gets calculate during the week, so that these resets don't cost them much or anything. Sometimes it feels like the usage just disappears with nothing to show for it.

> OpenAI obviously is doing it to gain mindshare but they're burning money before their IPO and can't sustain these resets.

We don’t know that. If they have already paid for the hardware and it is not running 100%, and customers would not pay to get reset, they don’t really lose money.

I personally believe that they are buying time until they can make current gov. to bring-in a legislation that targets Chinese models one way or the other.

And now Kimi-K3 has caused a new Deepseek moment.. but this time, the only horizon that was still untouched: frontier performance.

Expect escalation of things from now on.

They also got a lot of goodwill by letting people use codex subscription in any harness they want, anywhere they want

I think it's mostly to spread hype for the new models. Sol can be ridiculously long running, even without /goal so it can run for 12hr+ on a problem with defined and verifiable output. So it's a good way for people to get hyped about the capabilities without worrying about usage limits.

It's a pretty smart move. If they weren't doing this, I would have simply switched back to DeepSeek, Mimo, GLM, etc. (and maybe K3) whenever I hit my usage limits, as was my normal custom.

They're performing the resets when the cluster load goes down and they have sufficient capacity to handle the reset.

This feature has just been added!

There is a concrete expiry date listed for each reset in the codex desktop app.

30 days

When did that happen with Codex? I thought that was a Claude Code thing

I think it's perfectly normal for errors to occur given the pace of development.

Retroactive refunds are certainly better, but I think OpenAI is much more customer-focused in this regard than Anthropic.