I use a small percentage of my Pro token quota to find better deals on hotels, rental cars, discount codes for large purchases, etc. Have it automatically scanning credit card points promos and flight deals via various APIs too. Easily pays for itself in that regard. Over $1k, I'm not sure it would break even anymore.
Cosmo Kramer: It's a write-off for them.
Jerry: How is it a write-off?
Cosmo Kramer: They just write it off.
Jerry: Write it off what?
Cosmo Kramer: Jerry, all these big companies, they write off everything.
Jerry: You don't even know what a write-off is.
Cosmo Kramer: Do you?
Jerry: No, I don't.
Cosmo Kramer: But they do, and they're the ones writing it off.
Not 100% in the first year (where I live). By the time it’s depreciated, it’s time to replace. Subscriptions are 100% deductible in the year you pay them.
How would that work? It's not simple to set a price, when you have a fixed asset. Seems more complicated to set price on a product (ChatGPT), where the product is just a shell for something else entirely (the model), that changes by the month.
On what tokens? The ones of today's model, or tomorrow's?
The API makes it simple, yes. They make you pay per token per model. A new model has different prices. Not being able to do that is what makes it not so very simple.
As a consumer paying $20 now to chatGPT where i hit those gauntlet usage messages then using Muse a lot and not receiving any such messages. Im questioning my $20 a month GPT subscription and my $5 a month Gemini (Google has enough money as Meta to make it free) subscription.
Overall Ai is doing all the software development now and we the consumers are paying for this while many are out of work and if you are out of work the US tech job market is brutal!
Ai needs to be free to use without limits and in time Im guessing it will be. That could be Meta and or another who accomplishes this. Whoever does wins this race.
It's strange if that ends up to be this way. It's hard to see why would anybody subscribe to the $200 - or even the $500 for the sole benefit of using ultrafast (depleted probably within hours).
Seriously, even the "Learn more about limits" link just takes you to the posted page where it literally says nothing about limits or usage. Not even x20 or x10 or whatever, just zero information on what "more usage" actually means. I don't understand why they can't provide how much token usage you get.
Literally just downgraded to the $20 a month one, as they're not grandfathering in the old plans. I don't think this is going to work the way they thought.
I’ll be doing the same as you 10/26 when our grandfather limits end, in addition to downgrading my Anthropic sub. I guess I’m shopping for hardware and open weight AIs now.
> New subscriptions that aren’t eligible for grandfathering include a lower usage allowance than previously offered with Pro 200 to reflect our increasingly efficient models.
Everyone happily paid the way for the behemoth that squeezed everyone out of competition and then started recoup the ROI aggressively, and no one recognizes oneself as a part of the problem, just like in the current AI pandemonium.
I was happy to take VC money to uber to parts of town I wouldn't have otherwise for that period of time. Only my local bar lost out on selling me like maybe 20 drinks over the year or so.
I'll be the first to bemoan these companies but there are many, many locations where even with today's higher prices and level of service they come out way cheaper and way more reliable than incumbents.
It's getting shitty, I agree, but still feels like a net improvement to where we were over a decade ago.
Yeah, with Opus 5.5, my impression is that the OpenAI $200 plan was already worse than the Anthropic $200 plan. Sol 6.1 might have tilted the balance in OpenAI's favor again, but these changes more than negate that.
I'll cancel my OpenAI plan after this billing cycle. Let's see how this plays out.
I wonder how this will go with the new releases like MiMo-v2.6-Pro or future GLM, DeepSeek, and Kimi releases around the corner. Once planned with Fable or Astra, the build process isn't much different with the open weight models from my usage.
you can pretty much run frontier-level models locally now with GLM 5.3 Flash and Qwen 3.8 Flash Next so this pricing feels very geared towards squeezing people who are still dependent on cloud subscription models only and don't know how to divert grunt work to cheaper models automatically with the 'smarter' frontier models acting as orchestrators and refinement
Could hire real people in India for this cost - Fresh college under grads are paid like $250 per month or less for a month working 9hours x 5.5 to 6 days.
Hmm, a super-programmer familiar with almost every framework in existence that can ship entire new apps in a day and collaborate with you realtime, or paying to train someone in a different time zone for some likely subpar work?
At this rate I expect end-to-end work getting done, no babysitting whatsoever. The agents should literally be calling me and talking about their problem if something needs my input.
> If your Pro 200 subscription was active at the eligibility cutoff or during the seven days before it, you’re eligible to keep your previous included usage allowance through Oct 29, 2026 while you have an active Pro 200 subscription.
> After that date, your subscription will move to the lower included usage allowance. Your subscription price stays at $200/month.
Imagine buying this shit for a year, upfront, and then having to deal with this. Nice.
This day may go down in history as one of the worst in OpenAI's history. It's so bad it's funny.
OpenAI announced they built clones of apps that already exist, codex for normal people, and 2.5x price increase for the product that is already getting decimated by competition.
What is happening at OpenAI?
I can't imagine OpenAI not backtracking this unless Anthropic does the price hike too.
Price escalation and reduction in features just leaves a horrible taste in my mouth as a consumer, and moreso, a human.
I want them to succeed as a business so I can use their products, but trust does erode over time. I wish they would realize this and value their customer more.
Yea, the $500 package is likely a rebranded $200 from a few months ago. Keep in mind all providers have been cutting token allowances for their subs for a few months now
Yeah sorry I misunderstood you.
And yeah I bet that it is the cerebras collaboration although they didn't explicitly say this, maybe their own chip is fast enough for this inference speed? But I could only take a guess if its this or that
Yeah no. $200 was the top for me. Which doesn't super matter to me, I pretty much only use luna. I can't tell the difference between luna vs sol/astra other than those larger models do more ambitious things that I don't ask for.
The performance changes between models are almost static. Recently, GPT has shifted almost entirely toward token efficiency.
Astra is good at making 3D assets, but realistically its coding was not as good as GPT 5.6 Sol.
Have we reached a point where getting better at one thing means starting to get worse at another?
This announcement is the worst one yet.
Nerfing existing usage
A more expensive top-tier plan
Showing off how slow Dots is right in the demo video
And GPT Space—are they telling us to use it instead of Notion or what...
Overall, throughout the entire announcement, everything OpenAI wants me to look forward to just leaves me, as a developer, disappointed.
Yep. It's hitting limits because models have effectively reached the logical endpoint of how "precise" their outputs can become. In order to overcome this they need to either a) get excellent at outputting and working with massive codebases (10+ million loc) since a singificant % of systems simply cannot be done in less than that (the language itself isn't expressive enough) or b) reach the next stage of intelligence where the models are somehow able to output sequences that can solve vast complex problems with a very narrow output. Option b isn't really possible (various statistical/computational fundamental limitations forbid it) and option A is _insanely_ expensive. Which is why all of the demos the frontier labs have been pushing out are either really impressive one shot demos where the model is able to take a concise input and produce something great on its own, or very long running internal sessions where they claim to do something vast with minimal oversight (NavierStokes, C compiler, Bun rewrite).
At this price points (including the reduction of usage on the Pro 200 and Pro 100 subscriptions), more people will be looking buying their own hardware and running with open weight AIs. $6000 per year is the threshold
Decent(-ish), but doesn't provide good value for the money, sadly. They still refuse to tell you exactly how much usage you're paying for. My original suggestion a few months ago was to provide a $500/$1000 tier but make it _unlimited_, which this doesn't appear to be.
Very cool how I'm supposed to feel good about paying $200/mo (a VERY expensive subscription!) for months, just to be told that in the near (but not immediate) future this subscription will lose half of its value and then I'll have to pay 2.5x as much to get 1.25x the value I used to get.
Am I just out of touch with our times if I think hand-wavey usage limits on a 500$ subscription is crazy?
It filter to people/organizations who have a hand-wavey relationship with $500.
I use a small percentage of my Pro token quota to find better deals on hotels, rental cars, discount codes for large purchases, etc. Have it automatically scanning credit card points promos and flight deals via various APIs too. Easily pays for itself in that regard. Over $1k, I'm not sure it would break even anymore.
Fuck, that's smart. Thank you for the idea.
Usage is sus but 500 minus tax write off to even a one person business that makes 100-500k yr. not crazy at all - kinda expected.
Worth it? Huge YMMV. I'd expect a $1000 sub next.
You can also right off actual $6000 hardware
Not 100% in the first year (where I live). By the time it’s depreciated, it’s time to replace. Subscriptions are 100% deductible in the year you pay them.
Im surprised some countries dont have consumer protection laws that would force them to be more explicit/accurate
but to do accurate billing, they would have to tell us when our usage triggers a guard rail and causes our query to be rerouted or denied.
And if they do that accurately, there's no stopping someone from mapping the guard rails.
Clearly that outcome is more uncomfortable than what we have now, which is enough people seem to be okay with vague billing.
First, I wouldn't put it past OpenAI to charge even if your prompt triggered the guardrails, transferring responsibility is their style.
Second, how does vague billing prevent someone from mapping the "I'm sorry, I can't help you with that" style of responses to map the guardrails?
Move fast and bill things
or they could simply remove the annoying guardrails
How would that work? It's not simple to set a price, when you have a fixed asset. Seems more complicated to set price on a product (ChatGPT), where the product is just a shell for something else entirely (the model), that changes by the month.
Very easily, they do it for their API. You just set a limit in tokens.
On what tokens? The ones of today's model, or tomorrow's?
The API makes it simple, yes. They make you pay per token per model. A new model has different prices. Not being able to do that is what makes it not so very simple.
Maybe they're trying to milk as much as they can before open-weight models cause the prices to come crashing down?
As a consumer paying $20 now to chatGPT where i hit those gauntlet usage messages then using Muse a lot and not receiving any such messages. Im questioning my $20 a month GPT subscription and my $5 a month Gemini (Google has enough money as Meta to make it free) subscription.
Overall Ai is doing all the software development now and we the consumers are paying for this while many are out of work and if you are out of work the US tech job market is brutal!
Ai needs to be free to use without limits and in time Im guessing it will be. That could be Meta and or another who accomplishes this. Whoever does wins this race.
Yes, you are. All of this was obvious from the start. The prices will go up way more soon (tm)
Yeah it's wild that they don't outright state the specific usage limits.
Just trust me bro accounting.
Remember when people talked abouthow these models are similar to gambling. Well, might be preetty much loot box logix crossed with DLC
Horrid value.
Previous $200 tier: 20x pro value
New $200 tier: 10x
New $500 tier: 25x, plus ultrafast mode but it drains 6x the cost (!!)
Where do you see that? I can't even see the $500 plan on their pricing page.
The $500 25x was mentioned by Altman on the livestream
The $200 cut to 10x was announced only formally on Twitter
However none of their product pages mention concrete #s, so I guess value will only degrade more over time.
Thanks...
It's strange if that ends up to be this way. It's hard to see why would anybody subscribe to the $200 - or even the $500 for the sole benefit of using ultrafast (depleted probably within hours).
The pricing is on the linked page but does not appear to be selectable in ChatGPT UI yet.
The plan page just says "More usage" it doesn't say any 10x multiplier. Literally just says "More" than plus.
Seriously, even the "Learn more about limits" link just takes you to the posted page where it literally says nothing about limits or usage. Not even x20 or x10 or whatever, just zero information on what "more usage" actually means. I don't understand why they can't provide how much token usage you get.
Yes they used to show an estimate of how many messages you can send in each tier and for each model.
But past few weeks, I don’t think it was accurate since my usage was going down faster. They nerfed the usage and now aren’t even giving estimates.
Literally just downgraded to the $20 a month one, as they're not grandfathering in the old plans. I don't think this is going to work the way they thought.
I’ll be doing the same as you 10/26 when our grandfather limits end, in addition to downgrading my Anthropic sub. I guess I’m shopping for hardware and open weight AIs now.
I checked it's actually awful lol
> New subscriptions that aren’t eligible for grandfathering include a lower usage allowance than previously offered with Pro 200 to reflect our increasingly efficient models.
actual lol
> New rider trips include a higher base fare than previously offered with Uber to reflect our increasingly efficient trip routing. - Travis K.
Enjoy the Lyft v Uber era while it lasts!
lol yeah, I opened the Uber app the other day to $100 fares, remembered a time when I could get anywhere in SF at any time for like $5.
$5 flat group rides, $8 flat single rides anywhere in the city. The glory days.
Everyone happily paid the way for the behemoth that squeezed everyone out of competition and then started recoup the ROI aggressively, and no one recognizes oneself as a part of the problem, just like in the current AI pandemonium.
I was happy to take VC money to uber to parts of town I wouldn't have otherwise for that period of time. Only my local bar lost out on selling me like maybe 20 drinks over the year or so.
I'll be the first to bemoan these companies but there are many, many locations where even with today's higher prices and level of service they come out way cheaper and way more reliable than incumbents.
It's getting shitty, I agree, but still feels like a net improvement to where we were over a decade ago.
Oh don't worry, you get to keep your higher usage allowance for one month. hahahaha
I guess they got tired of people saying Codex plans were a better value than Claude Code.
Tired of their investors saying that
they were! now they are not
Conveniently Opus 5.5 is actually usable vs 5.
But I'm still for investing in local models.
Yeah, with Opus 5.5, my impression is that the OpenAI $200 plan was already worse than the Anthropic $200 plan. Sol 6.1 might have tilted the balance in OpenAI's favor again, but these changes more than negate that.
I'll cancel my OpenAI plan after this billing cycle. Let's see how this plays out.
> See the pricing page for a comparison of the features and usage included in each plan.
Does anyone see the usage limits? I don't see anything indicating how much usage is in each tier.
I checked:
* https://chatgpt.com/pricing/
* https://help.openai.com/en/articles/9793128-about-chatgpt-pr...
* https://chatgpt.com/#pricing (signed in plan upgrade page)
These days it's all like "$200 plan has more usage than $100 plan" and "$500 plan has even more usage".
It’s about time for some laws against unpredictable pricing like that.
To those balking at the price: this is just the beginning.
Subscription pricing will eventually compete with labor costs
I wonder how this will go with the new releases like MiMo-v2.6-Pro or future GLM, DeepSeek, and Kimi releases around the corner. Once planned with Fable or Astra, the build process isn't much different with the open weight models from my usage.
you can pretty much run frontier-level models locally now with GLM 5.3 Flash and Qwen 3.8 Flash Next so this pricing feels very geared towards squeezing people who are still dependent on cloud subscription models only and don't know how to divert grunt work to cheaper models automatically with the 'smarter' frontier models acting as orchestrators and refinement
https://this.os.isfine.org/blog/posts/chinese-flash-models-a...
Could hire real people in India for this cost - Fresh college under grads are paid like $250 per month or less for a month working 9hours x 5.5 to 6 days.
Could you hire hundreds of them over long running tasks for that fixed price to work every day and night for a month?
But they're slower and don't work in my timezone. The value to something like this isn't the ability to do work, but the ability to quickly iterate.
Hmm, a super-programmer familiar with almost every framework in existence that can ship entire new apps in a day and collaborate with you realtime, or paying to train someone in a different time zone for some likely subpar work?
My favorite comparison was that we're basically buying a GTX 1070 Ti every month, except instead of a GPU we get some extra productivity.
It's funny how, just before that, they tactfully cut the limits on the $200 subscription.
Been doing it slowly for months week over week allowances have been getting cut by 5-10% consistently
People complained about burning their weekly usage in a day on Pro 200, so OpenAI's response to that is an even more expensive plan?
Way easier to do with Anthropic plans, but I guess subsidized AI is over? Didn’t think it was happening this soon.
For a second I was worried I somehow missed 490 versions.
Lowering usage limits, increasing price.
This is just the beginning. Migrate to open models ASAP.
What a strange idea to name your product after it's price.
On the plus side, it'll be easier to compare these new plans than what we have with all the models.
It looks like they have literally chosen the most straightforward sounding plan names they could think of.
At this rate I expect end-to-end work getting done, no babysitting whatsoever. The agents should literally be calling me and talking about their problem if something needs my input.
We’re getting pretty close to that being the normal process at this point
I know people who have been at this as stage since around 4o with their own custom systems
Me too, I'd say we are about two weeks away
Great naming scheme that doesn’t even survive exchange rates, or future price changes.
Venture capital doesn't want to pay for our tokens anymore, it seems
dangit, I've been loving the subsidies.
> If your Pro 200 subscription was active at the eligibility cutoff or during the seven days before it, you’re eligible to keep your previous included usage allowance through Oct 29, 2026 while you have an active Pro 200 subscription.
> After that date, your subscription will move to the lower included usage allowance. Your subscription price stays at $200/month.
Imagine buying this shit for a year, upfront, and then having to deal with this. Nice.
You can't, Pro is only available with monthly payment.
No actual mention of the actual 'cutoff date' anywhere to be seen. Was it today? Yesterday? 6 months ago?
This day may go down in history as one of the worst in OpenAI's history. It's so bad it's funny.
OpenAI announced they built clones of apps that already exist, codex for normal people, and 2.5x price increase for the product that is already getting decimated by competition.
What is happening at OpenAI?
I can't imagine OpenAI not backtracking this unless Anthropic does the price hike too.
> I can't imagine OpenAI not backtracking this unless Anthropic does the price hike too.
Same here, but I’m worried it’s the latter. We’ll probably see an Anthropic price hike in a couple of weeks.
The $20/month subscription is an excellent deal. I wonder what these whales are doing with all those tokens?
I mean, other than Yegge. He seems like a pretty extreme outlier?
Found a whale: "It took OpenAI's agents 130 billion tokens to crack a 90-year-old math problem" https://www.businessinsider.com/openai-math-problem-solved-t...
DeepSeek 4.1 Flash is good enough for me thanks.
Given that I burn through 2 x $200 subscriptions, I really hope Sol 6.1 is indeed as good as Astra (￣ー￣；
Pure scam pricing.
Why would I pay $500 now when the $200 features were just rug-pulled? I mean, Pro 1000 is on the horizon at this point, no?
Price escalation and reduction in features just leaves a horrible taste in my mouth as a consumer, and moreso, a human.
I want them to succeed as a business so I can use their products, but trust does erode over time. I wish they would realize this and value their customer more.
Yea, the $500 package is likely a rebranded $200 from a few months ago. Keep in mind all providers have been cutting token allowances for their subs for a few months now
Is this Astra running on Cerebras?
Almost certainly yes. The value prop here is the 8x speed. This is incredibly valuable to many many builders.
No this is just the 500$ plan. They offer "ultrafast" now which is ~300tps which is 8x the speed for 6x the price.
I think 300 tps might be roughly what an astra sized model would be at if running on Cerebras hardware no?
Or are they simply lower the batch sizes on whatever hardware they are currently running?
Yeah sorry I misunderstood you. And yeah I bet that it is the cerebras collaboration although they didn't explicitly say this, maybe their own chip is fast enough for this inference speed? But I could only take a guess if its this or that
Yes that's their new ultrafast tier which is only available on this plan
Yeah no. $200 was the top for me. Which doesn't super matter to me, I pretty much only use luna. I can't tell the difference between luna vs sol/astra other than those larger models do more ambitious things that I don't ask for.
Is AI development hitting its limits?
The performance changes between models are almost static. Recently, GPT has shifted almost entirely toward token efficiency.
Astra is good at making 3D assets, but realistically its coding was not as good as GPT 5.6 Sol.
Have we reached a point where getting better at one thing means starting to get worse at another?
This announcement is the worst one yet.
Nerfing existing usage A more expensive top-tier plan Showing off how slow Dots is right in the demo video And GPT Space—are they telling us to use it instead of Notion or what...
Overall, throughout the entire announcement, everything OpenAI wants me to look forward to just leaves me, as a developer, disappointed.
Yep. It's hitting limits because models have effectively reached the logical endpoint of how "precise" their outputs can become. In order to overcome this they need to either a) get excellent at outputting and working with massive codebases (10+ million loc) since a singificant % of systems simply cannot be done in less than that (the language itself isn't expressive enough) or b) reach the next stage of intelligence where the models are somehow able to output sequences that can solve vast complex problems with a very narrow output. Option b isn't really possible (various statistical/computational fundamental limitations forbid it) and option A is _insanely_ expensive. Which is why all of the demos the frontier labs have been pushing out are either really impressive one shot demos where the model is able to take a concise input and produce something great on its own, or very long running internal sessions where they claim to do something vast with minimal oversight (NavierStokes, C compiler, Bun rewrite).
At this price points (including the reduction of usage on the Pro 200 and Pro 100 subscriptions), more people will be looking buying their own hardware and running with open weight AIs. $6000 per year is the threshold
It's a bit more complicated than that though. Dollars per token is not the right metric to track.
Related:
Tomorrow we are re-opening the Pro $200 subscription
https://news.ycombinator.com/item?id=49889306
Decent(-ish), but doesn't provide good value for the money, sadly. They still refuse to tell you exactly how much usage you're paying for. My original suggestion a few months ago was to provide a $500/$1000 tier but make it _unlimited_, which this doesn't appear to be.
Very cool how I'm supposed to feel good about paying $200/mo (a VERY expensive subscription!) for months, just to be told that in the near (but not immediate) future this subscription will lose half of its value and then I'll have to pay 2.5x as much to get 1.25x the value I used to get.
Fuck that shit!